I saw something like 62 and I really started to panic. Was I falling sick? Did I get too stressed? Did I sleep the wrong way?

In reality, this didn't matter because the ring didn't even know me at all. It only had one night's worth of information on me. One. And I was out here treating it like gospel.

For those who have purchased a smart ring or smart watch only to be disappointed by the fact that it doesn't instantly understand them, this one is for you. The truth is your wearable isn't broken in the first week. It just doesn't know you yet.

Table of Contents

• Your First Few Days Are Basically Noise

  • What "Baseline" Actually Means Here

• How Wearables Actually Build That Baseline

• Why It Takes Weeks, Not Days

• The Cold Start Problem Isn't Unique to Wearables

• What This Actually Means for You

Your First Few Days Are Basically Noise

Whether it's an Oura, a WHOOP, an Ultrahuman, or a Garmin, each wearable operates under the assumption that it'll have a personal baseline for you, which is your norm. This includes metrics such as your resting heart rate during a typical day or night, HRV range, and body temperature prior to any unusual circumstances occurring.

The issue here lies in the fact that each data point will never be able to provide any information about your norms and baselines. All it does is show the results of your specific measurement. Is 45ms of your heart rate variability considered low for you or high for you? The device won't know until it sees another similar data point.

This is exactly why the readiness or recovery score during the first week of using a new device looks somewhat random to you. It's measuring you against a rough population average, or worse, guessing. Some brands are upfront about this, most just quietly show you a number and let you assume it means something solid.

What "Baseline" Actually Means Here

In terms of wearables, a baseline isn't an average. It's a dynamic range that's created from several days or nights and often comes with a standard deviation band. In other words, what you want to hear is not "your HRV is 45" but rather "your HRV is normally between 38 and 52, and 45 today is absolutely normal for you".

This range is exactly what we're talking about. Without the range, none of the numbers would be alarming or useless, as there would be nothing to compare it with.

And it's not only HRV. Your resting heart rate, skin temperature, respiratory rate and even your level of restlessness during nighttime – everything gets its own baseline, created independently, on its own timeline.

Skin temperature is the one that requires the smallest amount of data points for stabilization, as it's a narrow range for the vast majority of people. HRV, on the other hand, needs a lot more as it varies depending on stress, alcohol, load of workouts, state of illness, and many more factors.

How Wearables Actually Build That Baseline

In most consumer wearables, there's a variant of the exponentially weighted moving average or EWMA, used to establish your baseline over time. Recent nights matter more than older ones, but none is discarded outright.

It differs from regular averaging by allowing for gradual changes in your baseline to reflect your actual changes without a single off-night ruining the whole thing.

For instance, Oura openly talks about rolling baselines used in their calculations of skin temperature based on your recent history. That's also why the graph of temperature deviations becomes meaningful only after you've been wearing the ring for a while. WHOOP works similarly to build a baseline for HRV and resting heart rate until it decides it's time to produce truly personal recovery scores.

The math behind this isn't complicated, even if it sounds fancy. Any new reading contributes only a tiny bit to building up the baseline. Ten percent weight per day is probably a good number to think of, while ninety percent remains locked into everything that came before.

This way, one bad night, one good night, or one weird reading from a faulty sensor can't ruin the baseline single-handedly. It needs repeated data to be affected and it's the reason why it takes weeks and not days to build up, because a few days can't outweigh accumulated history.

And this is the whole point, but it also means your very first baseline estimate is basically wearing training wheels.

This approach is basically the cold start problem that you can see in recommendation systems or other machine learning models that need historic data in order to produce meaningful results.

In some cases, it may be implemented as a two-stage process on purpose. During the first stage, typically around the first two weeks, you're scored against a generic population model built from data of thousands of other users with similar age or profile. When your history has enough data points already, it silently switches you to scoring against yourself. There's no such thing as an explicit switch here. Instead, people simply notice that scores become more accurate.

Why It Takes Weeks, Not Days

Despite an intelligent weightage system, it's impossible to gather diverse data in just a few days. Your body doesn't go through all the normal parameters in three nights. A baseline should include:

• A regular work night and a rough sleep night

• A rest day and a day with an intense workout

• Normal stress and a stressful day at work

• In case of a menstrual cycle, ideally several stages of the period, since HRV and temperature fluctuate greatly in the course of the cycle

That's not something that happens in 72 hours. That's why two to four weeks are recommended by most wearable companies prior to interpreting the results for practical use, despite the fact that it may not be emphasized in bold letters on the box, since "wait a month for this thing to become useful" doesn't exactly sell.

I've felt this firsthand with cycle tracking specifically. My temperature baseline on any ring looks completely different depending on which week of my cycle it was built during. If the algorithm only had follicular phase data, my luteal phase numbers would look like a fever every single month. Give it a full cycle or two and suddenly the deviations actually mean something instead of triggering a false alarm every three weeks.

The Cold Start Problem Isn't Unique to Wearables

This is the problem recommendation engines face with their new users as well. There's no way for Netflix to provide any kind of meaningful recommendation based on the user watching a single movie. And a wearable device can’t do anything with one day of sleep analysis.

The solutions in both cases are pretty similar as well. The approach here is to rely on population data or some general defaults first and then slowly replace it with individual data over time, giving more weight to recent behavior than to older ones because lifestyles change and bodies change. That’s essentially the same approach but presented differently.

So if you're building something that touches biometric or behavioral data, this is an approach that should be accounted for from the very beginning rather than implemented later. Show users a confidence indicator instead of a clean, confident-looking number when the model barely has any history to work with. A recovery score of 61 looks equally authoritative on the second day as on the sixtieth one, despite the difference between the scores and the reason why. Being honest about that difference in the UI saves you a lot of confused support tickets down the line.

What This Actually Means for You

This article will help if you just bought a new ring or watch. Here's what you should do with it during the first month.

First, ignore your daily scores in week one. They aren't completely wrong, they're just based on almost nothing. You can use them the same way as the first impression by someone who doesn't know you.

Next, write down some context if your app provides that opportunity. Add tags like alcohol, poor sleep, travel, illness, workouts, and so on. The sooner the algorithm learns different information about you, the sooner it learns the true baseline.

Also, wait at least two to three weeks to assess the accuracy of the results. It's roughly the period when the algorithm receives enough information to compare rather than guess.

Finally, track a whole cycle or two if you track your period. Otherwise, the algorithm gets only part of your cycle and gives you inaccurate information. That's the reason why sometimes people get an alert about "low readiness" in the luteal phase of a perfectly normal cycle.

I still check my ring every day, but I stop taking the results from week one seriously for a long time. The data starts to be truly helpful after I've given enough time to the algorithm to learn all the details of my life, including both good and bad periods. And until that moment, it's not telling lies to you, it just hasn't learned enough yet.

So the next time when your brand new device shows a score that seems wrong or even insulting, remember it isn't judging you. It just has the sample size of one and tries to work with it the best way. Give it the weeks it actually needs and it'll start telling you something real instead of just making an educated guess and hoping for the best.

We just published a talk on the freeCodeCamp Spanish YouTube channel about how to leverage freeCodeCamp's free learning resources and community to start your career in technology. You’ll learn how to find these resources and the core concepts that you need to know to start contributing to open source.

If you have Spanish-speaking friends, you're welcome to share the Spanish version of this article with them.

This talk was presented by Estefania (me!). I'm part of the freeCodeCamp team. I develop educational content and manage the freeCodeCamp Spanish YouTube channel. I love helping others learn and grow professionally. I gave this virtual talk for the Escuela Superior Politécnica del Litoral, located in Guayaquil, Ecuador.

Why Consider a Career in Tech?

Before we dive into the content of the talk, let's see what careers in technology involve, and why you should start now if having a career in this field is your goal.

Technology careers can be very interesting because they give you the opportunity to solve real-world problems. Being part of the technology field means that you'll have the chance to help shape the future of humanity. By joining the freeCodeCamp community, you'll find the support and the practical tools you need to get started.

Learning to code and contributing to open source projects, like freeCodeCamp, are two fundamental steps for reaching your goal. Contributing to open source projects develops your technical skills and connects you with a global network of developers who share your interests and goals. By contributing to open source, you’ll gain hands-on experience, improve your portfolio, and increase your chances of landing your first job.

What You'll Learn During the Talk

Great. Let's see what you’ll learn during the talk:

• The story and mission of freeCodeCamp.org.

• How to use freeCodeCamp’s free learning resources to learn programming.

• freeCodeCamp's certifications.

• freeCodeCamp's daily coding challenges.

• freeCodeCamp's catalog and forums.

• freeCodeCamp's YouTube channels with full courses.

• freeCodeCamp's publication.

• Additional free learning resources provided by freeCodeCamp.

• How to contribute to open source projects and why this is important.

• Common terminology used in open source projects.

• Personal tips for getting started in your first job.

• How to join the freeCodeCamp community.

By the end of the talk, you’ll know how to find and leverage freeCodeCamp's free learning resources to learn programming and start your career.

Talk on YouTube

Check out the talk on the freeCodeCamp Spanish YouTube channel:

✍️ Talk presented by Estefania Cassingena Navone.

• YouTube: https://www.youtube.com/@freecodecampes

Collaborating with:

• Canal del Capítulo (IEEE Computer Society ESPOL): https://www.youtube.com/@ieeeespolcomputersociety7497

• Canal de la Rama Estudiantil (IEEE ESPOL): https://www.youtube.com/@ramaestudiantilieee-espol84

From fraud prevention to real-time payments and smart investing, AI is unlocking major opportunities across finance. Machine learning models help identify suspicious activity faster than ever before, while also enabling hyper-personalized customer experiences. AI-driven payment systems improve transaction speed, reduce friction, and make financial services more accessible worldwide.

In investing and trading, predictive analytics and NLP help firms uncover market insights, assess risk, and automate decision-making. From hedge funds to robo-advisors, AI is enhancing performance and democratizing access to financial tools.

Globally, AI is also strengthening cross-border collaboration and compliance. Through APIs, real-time data sharing, and regulatory tech, financial institutions are creating more transparent and agile systems that operate across jurisdictions.

This handbook explores how AI is driving the next era of finance. Whether you're a bank executive, fintech innovator, or policy leader, you’ll find practical insights and tools to guide your organization into a smarter, data-driven future.

“You are not going to lose your job to AI, but you are going to lose your job to a developer who uses AI.”

– Jensen Huang, CEO @NVIDIA

Table of Contents:

1. Chapter 1: Why AI in Finance Is a Necessity – Not Just Hype

2. Chapter 2: AI in Finance Today – Where Are We in AI and Innovation?

3. Chapter 3: Case Studies of AI in FinTech – Global Use Cases and Case Studies of AI in Finance

4. Chapter 4: The Role of Data in Finance – Infrastructure, Warehousing, and Security

5. Chapter 5: The Science Behind the Models – ML, NLP, and Predictive Analytics

6. Chapter 6: Training the Workforce – Upskilling Executives, Technical, and Non-Technical Teams in FinTech

7. Chapter 7: Resources for Finance Executives – AI Education & Enablement in Finance: Workshops, Tools, Services, and Training Resources

You can download the PDF Version of the eBook here.

And you can also listen to this handbook as a podcast here:

Chapter 1: Why AI in Finance Is a Necessity – Not Just Hype

The financial sector has long prided itself on being ahead of the curve when it comes to adopting new technologies. From early mainframe systems to real-time trading platforms, banks, hedge funds, and payment providers have historically been quick to embrace tools that promise greater speed, efficiency, and insight.

But the world has changed – and fast.

Today, Artificial Intelligence (AI) and data-driven technologies are redefining what innovation means in finance. From predictive risk modeling to hyper-personalized customer experiences, AI isn’t a buzzword or a future luxury. It’s a present-day requirement for survival.

The Innovation Gap: Perception vs. Reality

It may surprise you that even in some of the world’s most digitally advanced regions, many financial institutions still rely heavily on legacy systems. Core banking infrastructure often runs on outdated technologies. Manual compliance checks, fragmented data storage, and lack of real-time analytics are still common.

In countries with strong financial histories, legacy often gets in the way of progress. While fintech startups sprint ahead with cloud-native, AI-first approaches, traditional banks and insurers are struggling to digitize core services, let alone lead with data.

This isn’t just a minor gap – it’s a growing risk. Institutions that delay digital transformation fall behind not only in customer service but in risk mitigation, fraud prevention, and investment performance.

Where Innovation Is Needed

AI isn’t a one-size-fits-all solution. But it offers specific, actionable advantages across nearly every domain of finance:

• Retail Banking: AI improves customer service, personalizes offerings, detects fraud in real-time, and enables better credit decisions using alternative data.

• Investment & Asset Management: Predictive analytics help portfolio managers spot trends early. Robo-advisors offer scalable, custom investment advice. NLP tools turn earnings calls and market chatter into structured insight.

• Payments & Fintech: Machine learning models reduce fraud, optimize payment routing, and improve KYC/AML compliance with far greater accuracy.

• Insurance & Risk: AI models assess risk in real-time, automate underwriting, and help insurers respond to claims with minimal manual effort.

• Trading & Hedge Funds: From quant strategies using reinforcement learning to sentiment-based trading algorithms, AI has already reshaped trading floors.

• Compliance & Security: Natural Language Processing (NLP) automates the review of regulatory documents. Anomaly detection finds suspicious transactions that human analysts might miss.

In short: AI is not a tool to consider "someday." It’s an operational backbone for today and tomorrow.

It’s About ROI – Not Just Technology

With every AI buzzword, there comes hype – and with hype, hesitation. This is healthy. Financial leaders need to see measurable ROI, not just a list of features.

Smart AI adoption focuses on:

• Solving real business problems (for example, reducing loan processing time by 60%)

• Improving customer KPIs (for example, 20% higher retention from personalized financial advice)

• Cutting operational costs (for example, automating reconciliation processes)

• Enhancing security and compliance in increasingly hostile threat environments

This handbook is about moving past the hype and into real value.

Who Should Read This Handbook

This is a handbook written for decision-makers – executives, investors, and operators who shape the future of financial services:

• Bank executives and managers who want to transform operations and customer experience

• Fintech founders and product teams building next-gen platforms

• CTOs and CIOs tasked with modernizing infrastructure

• Investors – VCs, PEs, GPs, LPs – looking to evaluate scalable fintech and AI plays

• Leaders in asset management, hedge funds, and trading who want a performance edge

• Insurance and payment companies navigating digital acceleration

What to Expect

This handbook dives deep into how AI and data are being applied across the financial world – not in theory, but in practice. We'll explore global case studies from Singapore to New York, Tokyo to Amsterdam that show exactly how leading firms are deploying AI to solve real-world challenges.

We’ll break down the ecosystem into the most relevant financial verticals and explain:

• What problems AI solves

• How data infrastructure plays a role

• What tools and platforms are available

• How organizations can upskill their teams

• What successful case studies teach us

By the end of this handbook, you’ll walk away with a roadmap – not just for “adopting AI,” but for building a sustainable, data-driven financial institution that stays ahead of the curve.

Chapter 2: AI in Finance Today — Where Are We in AI and Innovation?

At its core, finance is the science and business of managing money – how it’s earned, saved, invested, insured, borrowed, and spent. That definition hasn’t changed. But the methods, expectations, and technologies that drive modern finance have radically transformed.

In today’s financial ecosystem, institutions are no longer judged solely on interest rates or product offerings. Instead, they are measured by:

• How fast they can deliver services

• How well they personalize customer experiences

• How securely they protect data and infrastructure

• How intelligently they manage risk and capital allocation

And most importantly, by how effectively they use data.

Finance in 2025: Data-Centric and AI-Driven

Every financial activity – be it a retail transaction, a cross-border payment, an IPO, or a wealth management advisory session – generates a digital footprint. What sets the leaders apart is how well they can capture, structure, analyze, and act on that data.

AI is the natural engine of this transformation. But today, we’re at a mixed adoption stage globally.

Where Finance Is Excelling in AI

Many large financial players have already implemented AI with impressive results. Here are a few standout areas:

• Fraud Detection and Risk Management: AI models can now detect fraud in milliseconds by analyzing real-time patterns and anomalies (for example, Mastercard and Visa use ML to detect fraudulent transactions before they’re completed).

• Algorithmic and Quantitative Trading: Hedge funds like Renaissance Technologies and Two Sigma use machine learning for predictive modeling based on vast data sources, including alternative data like satellite imagery.

• Robo-Advisors and Personal Finance: Platforms like Betterment and Wealthfront use AI to provide automated, personalized investment strategies at scale.

• Customer Service: Chatbots and AI-powered assistants are now handling millions of interactions across banks like Bank of America (Erica) and HSBC, significantly reducing customer support costs.

These are just the beginning. In many of these cases, AI has not just improved performance – it has become a core competitive advantage.

Where the Gaps Are

Despite high-profile innovation, many financial institutions – especially traditional banks and insurers in Western Europe, Southeast Asia, and Latin America – are lagging behind.

Common challenges include:

• Legacy Core Systems: Older, monolithic infrastructures make data integration and automation difficult.

• Siloed Data: Without centralized data warehouses or lakes, advanced AI modeling is almost impossible.

• Shortage of AI Talent: Many banks lack in-house AI engineers or data scientists, leading to reliance on generic third-party tools.

• Regulatory Fear: Concerns over compliance and data privacy (GDPR, AML, Basel III) often slow down innovation, even when AI can help meet those very obligations.

A 2023 report by the World Economic Forum noted that while 85% of financial executives see AI as “essential” to future growth, fewer than 35% have deployed it at scale within core operations.

This means we are still in the early innings – especially for those outside of major innovation hubs like New York, London, or Hong Kong.

Finance Is Becoming Fintech by Default

One important shift: the line between traditional finance and fintech is vanishing.

Any company that provides financial services must now think like a tech company. This includes retail banks, wealth managers, insurers, private equity firms, and central banks. Whether they like it or not, they are becoming data companies.

• Payments are being reinvented by APIs and machine learning optimization (Stripe, Adyen, Square).

• Lending is now algorithmic, with startups like Upstart and Kabbage approving loans in seconds using AI-based credit scoring.

• Investment analysis is real-time, with platforms scanning global news, earnings reports, and social media sentiment 24/7.

• Insurtechs are pricing risk more accurately than ever with real-time data from connected devices and behavioral scoring.

Legacy institutions that resist this shift risk being leapfrogged by more agile, AI-first challengers.

The Global Landscape: An Uneven Map

Innovation levels vary widely across regions:

• United States: Leading in AI-driven trading, wealth tech, and regtech. Heavy investment in AI research and startup ecosystems.

• United Kingdom: Strong fintech sector in London, but traditional banks remain cautious. Regulation-friendly for experimentation (for example, FCA sandbox).

• Netherlands & Germany: Wealth of talent and infrastructure, but legacy banking institutions are slow to adapt AI internally.

• Singapore & Hong Kong: Government-backed innovation hubs, strong adoption in wealth management and regulatory tech.

• China: AI-first approach in consumer finance and mobile payments, led by Ant Group and Tencent.

• Canada & Australia: Focused on ethical AI and compliance automation. Slower in retail innovation but strong in institutional tech.

• Japan: Conservative innovation pace in traditional banks, but increasing AI use in investment and manufacturing finance.

This variance opens the door for learning across borders – and for competitive advantage in under-served regions.

Finance today is not just about managing capital. It's about managing data, speed, trust, and intelligence. AI is no longer the edge. It is becoming the foundation.

In the next section, we’ll go beyond definitions and into real-world examples: How are top institutions – from Goldman Sachs to Revolut to Ant Financial – applying AI in ways that are changing the game.

Chapter 3: Global Use Cases and Case Studies of AI in Finance

AI is no longer experimental in finance – it's operational. From Wall Street to Shanghai, leading institutions are deploying machine learning, natural language processing (NLP), and generative AI not just to optimize processes but to redefine them.

In this section, we explore real-world case studies of how AI is already transforming financial services across banking, investing, payments, compliance, and customer experience. These examples span a global spectrum – from the U.S. to Asia to Europe – offering a comprehensive view of how AI is being leveraged across different financial sectors worldwide.

JPMorgan Chase – COiN (Contract Intelligence Platform)

Country: United States
Function: Legal automation and document review
AI Applications: NLP and Machine Learning
Impact: Reduced 360,000 hours of manual review time

JPMorgan’s COiN (Contract Intelligence) platform is a pioneer in AI for legal and compliance processes. Using Natural Language Processing (NLP), COiN automates the review of legal documents, particularly complex credit agreements. This process, which used to take hundreds of thousands of hours of manual work, is now completed in a fraction of the time, significantly enhancing operational efficiency.

• Risk Analysis: COiN scans documents to identify key terms, obligations, and risks associated with legal contracts. This allows compliance officers to focus on the high-risk contracts and flag potential issues early on.

• Operational Cost Savings: The automation provided by COiN reduces reliance on manual labor and minimizes the risk of human error, ultimately saving the bank time and money.

• Compliance and Speed: COiN helps JPMorgan comply with complex regulatory requirements by making the review process quicker and more accurate, reducing compliance risk.

COiN is a clear example of how AI can disrupt back-office operations, providing banks and financial institutions with tools that significantly improve productivity and legal oversight.

BlackRock – Aladdin (Asset, Liability, Debt & Derivative Investment Network)

Country: United States (Global deployment)
Function: Risk management, portfolio construction, investment operations
AI Applications: Predictive analytics, real-time risk modeling
Impact: Powers ~$21 trillion in assets under management

Aladdin, BlackRock’s AI-powered risk management platform, is one of the most influential tools in the investment management space. Aladdin leverages predictive analytics and real-time data to help asset managers assess risk, build portfolios, and manage their investment operations.

• Scenario Analysis: Aladdin simulates various market scenarios (such as changes in interest rates or economic downturns) to help portfolio managers identify potential vulnerabilities and optimize portfolio performance accordingly.

• Market Prediction: Aladdin uses AI to forecast asset performance by analyzing both historical and real-time data, allowing asset managers to make data-driven decisions that improve returns while managing risk.

• Operational Risk: The platform can quickly identify potential gaps in the operational side of portfolio management, providing actionable insights to reduce risks.

Aladdin is used by financial institutions around the world, including large asset managers, insurers, and sovereign wealth funds. By licensing its technology, BlackRock has turned into not just an asset management firm, but a technology provider as well.

Here’s a BlackRock Aladdin overview if you want to read more.

Goldman Sachs – Marcus & AI-Powered Consumer Finance

Country: United States
Function: Consumer banking, digital lending
AI Applications: Behavioral analytics, NLP, personalization
Impact: Over $100B in deposits managed via AI-augmented digital channels

Goldman Sachs entered the consumer banking space with Marcus, a digital platform offering savings accounts and personal loans. Powered by AI, Marcus has revolutionized how the bank approaches credit decisioning, personalized financial advice, and customer onboarding.

• Credit Decisioning: Goldman Sachs uses AI to assess creditworthiness by analyzing alternative data sources, such as transaction history and social behavior, instead of just traditional credit scores. This allows Marcus to extend credit to a wider customer base, especially those underserved by traditional banks.

• Personalization: AI-driven algorithms create tailored financial solutions for individual customers, such as personalized savings plans or investment recommendations, enhancing user experience.

• Automated Onboarding: The AI engine speeds up the verification process, reducing manual input and allowing customers to open accounts in a matter of minutes, rather than days.

Goldman Sachs’ move into the digital consumer finance space underscores how even traditional investment banks can innovate and compete with fintech disruptors by leveraging AI to improve user experience and streamline operations.

You can read more about Marcus by Goldman Sachs if you’re curious.

Ant Group – AI for SuperApp Finance

Country: China
Function: Mobile payments, credit, insurance, wealth
AI Applications: Deep learning, behavior-based credit scoring, fraud detection
Impact: Over 1 billion users served by AI-driven services

Ant Group, the parent company of Alipay, integrates AI throughout its extensive ecosystem, offering mobile payments, credit, insurance, and wealth management services. The scale at which Ant operates – with over 1 billion users – makes its AI deployment incredibly sophisticated.

• Zhima Credit (Sesame Credit): This AI-powered credit scoring system uses behavioral data to evaluate creditworthiness. By analyzing transaction history, utility bill payments, and even social behavior, Ant Group can offer personalized loans and financial products to users who may lack traditional credit histories.

• Fraud Detection: Real-time anomaly detection systems continuously monitor billions of transactions to flag suspicious activity, preventing fraud before it happens. This has greatly improved trust in digital financial transactions, particularly in regions where traditional banking infrastructure is lacking.

• Smart Customer Support: Ant's NLP-powered chatbots resolve over 95% of customer queries autonomously, ensuring users receive timely assistance.

Ant Group’s AI-driven platform enables massive scalability and efficiency, allowing the company to offer an array of services without the need for extensive physical infrastructure.

Revolut – Real-Time Fraud Detection and Personalization

Country: United Kingdom
Function: Neobank, payments, FX, crypto
AI Applications: Real-time anomaly detection, personalization engines
Impact: 35M+ users, AI flags >95% of fraud in real time

Revolut uses AI extensively to enhance both customer experience and security across its neobanking platform. By leveraging machine learning, Revolut is able to detect fraud in real time and personalize financial services for each user.

• Fraud Detection: Revolut’s AI models analyze behavioral patterns – such as location, transaction frequency, and device fingerprinting – to identify potentially fraudulent activities in real time. This allows the system to immediately flag suspicious transactions, ensuring a high level of security for its global user base.

• Personalization: Revolut’s AI engine provides users with customized budgeting tips, spending insights, and even recommends financial products such as loans and insurance, based on individual transaction data.

• Scalability: Revolut’s AI stack is designed to handle the massive scale of over 35 million users spread across 200+ countries, all while maintaining high standards of personalization.

Revolut’s success lies in balancing cutting-edge AI with a streamlined, user-friendly experience, proving that AI is not just a tool for large banks but also for nimble fintech startups.

You can read more about Revolut’s AI-driven approach here.

Renaissance Technologies – Predictive Quant Trading

Country: United States
Function: Hedge fund
AI Applications: Machine learning, alternative data modeling, signal extraction
Impact: Arguably the most profitable quant firm in history

Renaissance Technologies, the legendary hedge fund, is known for its AI-powered and data-driven investment strategies. The firm employs some of the most advanced machine learning techniques and data models to predict price movements, gaining a significant edge in the market.

• Alternative Data Analysis: Renaissance uses unconventional data sources such as satellite imagery, weather data, and even social sentiment from social media platforms to build predictive models. For instance, they may analyze the number of cars in the parking lot of a retail chain using satellite images to forecast quarterly earnings.

• Machine Learning Models: Renaissance Technologies uses machine learning models to identify patterns and signals that human analysts may miss, making their trading decisions faster and more accurate.

• Consistent Returns: The firm’s flagship Medallion Fund has reportedly returned over 60% annually (net), a remarkable feat in the investment world, thanks to its reliance on AI to optimize every aspect of its trading strategy.

Renaissance’s success story is a perfect example of how AI, combined with alternative data, can produce extraordinary financial returns.

Generative AI for Internal Automation and Client Interaction

Used Globally
Function: Customer service, internal productivity, compliance
AI Applications: LLMs (like ChatGPT), GPT-powered copilots
Impact: Reduces response time, boosts compliance, increases advisor efficiency

Generative AI is being rapidly adopted across the finance industry for internal automation and client interaction. AI tools like ChatGPT and similar Large Language Models (LLMs) have found applications across multiple facets of financial institutions:

• Customer Service Automation: Banks and financial institutions are using generative AI to power chatbots and virtual assistants that handle common customer inquiries, reducing the need for human intervention and significantly improving response times.

• Internal Productivity: AI copilots, like those tested by Morgan Stanley and UBS, help financial advisors quickly retrieve research, analyze market trends, and generate custom reports. This allows advisors to focus on more valuable, higher-level tasks like client engagement.

• Compliance Assistance: Generative AI is also being deployed to automate risk documentation, summarize compliance reports, and assist in the generation of legal documents, ensuring that the vast array of regulatory requirements is met with greater accuracy and efficiency.

Here are some examples:

• Morgan Stanley uses OpenAI’s GPT to help financial advisors access research instantly.

• UBS is testing AI copilots to assist relationship managers and client-facing bankers.

• ING uses AI to streamline internal processes like writing credit memos and risk assessments.

Generative AI is transforming how financial firms deliver customer service, assist employees, and maintain compliance.

Chapter 4 - Data Management in Finance: Navigating Data Lakes, Real-Time Ingestion, Security, and Cloud Platforms

In the digital age, data has become the lifeblood of the financial industry. From risk management to customer service and predictive analytics, financial institutions are increasingly relying on vast amounts of data to make informed decisions.

But handling this data requires advanced infrastructure, as well as a deep understanding of how different technologies can be leveraged to optimize data usage.

In this section, we’ll explore the critical components of data management in finance, including data lakes vs. data warehouses, real-time data ingestion, data security and compliance, and the role of cloud platforms like AWS, GCP, and Azure in managing financial data.

Data Lakes vs. Data Warehouses: The Foundation of Financial Data Management

When dealing with large volumes of data, teams and companies must decide how best to store, manage, and utilize that data. This decision often comes down to two key technologies: data lakes and data warehouses. While they may seem similar, they serve different purposes and have distinct advantages depending on the needs of the organization.

Data Lakes: Flexible and Scalable for Big Data

A data lake is a centralized repository that allows financial institutions to store vast amounts of structured, semi-structured, and unstructured data at scale. The key advantage of a data lake is its flexibility – it can accommodate data from a variety of sources without requiring any preprocessing or transformation.

In finance, data lakes are ideal for storing massive datasets such as transaction logs, market data, social media feeds, and customer interactions. By consolidating this data in one place, organizations can perform exploratory data analysis, conduct advanced analytics, and implement machine learning models.

Advantages:

• Scalability: Data lakes can handle petabytes of data with ease.

• Cost-Effective: They are often built on low-cost storage solutions, which makes them a cost-effective way to store large amounts of data.

• Data Variety: They can store data in its raw form, including structured data (like customer demographics), semi-structured data (like transaction logs), and unstructured data (like customer service chat logs or social media feeds).

Challenges:

• Data Quality: Since data in a lake is often stored in its raw form, ensuring the quality of the data can be challenging.

• Data Governance: Proper governance frameworks need to be in place to manage who has access to the data, and how it can be used securely and ethically.

Data Warehouses: Structured and Optimized for Analytics

A data warehouse, on the other hand, is designed for structured data that is preprocessed and optimized for analytics. It usually stores historical data, transformed into a format that is easy to query and analyze. In financial institutions, data warehouses are used for business intelligence, reporting, and making strategic decisions based on historical trends.

Banks and asset management firms often rely on data warehouses for financial reporting, risk management, fraud detection, and compliance tracking. It allows them to access a clean and structured dataset that is ready for analysis.

Advantages:

• Performance: Data warehouses are highly optimized for complex queries and fast analytics.

• Data Integrity: The data stored in warehouses is usually cleaned and transformed, ensuring a high degree of accuracy and consistency.

• Business Intelligence: They support advanced business intelligence tools and reporting features, helping executives make informed decisions.

Challenges:

• Cost: Data warehouses typically require more expensive storage and computing resources due to their structured nature.

• Rigidity: Unlike data lakes, data warehouses are less flexible when it comes to accommodating unstructured data or rapidly changing datasets.

Real-Time Data Ingestion and Processing: The Importance of Speed in Finance

The ability to process real-time data has become a critical factor for success in modern financial services. Whether it's market trading, fraud detection, or customer support, financial institutions need to ingest and analyze data as it happens to make timely decisions and maintain competitive advantage.

Real-Time Data Ingestion

In the financial world, real-time data ingestion refers to the continuous flow of data from various sources (such as stock markets, credit card transactions, or social media) into a central system for immediate processing. For instance, banks must process millions of transactions every second to identify fraud or assess liquidity risk.

• Example: A trading algorithm that ingests live market data (price movements, order books, and so on) and adjusts trading strategies in real time, helping asset managers to react instantly to market conditions.

• Key Technologies: Real-time data ingestion typically uses streaming technologies such as Apache Kafka, AWS Kinesis, or Google Cloud Pub/Sub to process and route data to processing systems with minimal delay.

Real-Time Data Processing

Once data is ingested, it needs to be processed immediately to generate insights or trigger actions. For example, real-time fraud detection systems analyze each credit card transaction as it happens to determine whether it’s legitimate or fraudulent, using algorithms that monitor patterns and behaviors.

• Key Processing Technologies: In finance, streaming analytics platforms like Apache Flink or Google Dataflow are commonly used to handle real-time data. These platforms allow institutions to run complex analytics on data in motion, enabling them to identify risks, opportunities, or irregularities quickly.

Use Cases:

• Fraud Detection: Banks and payment processors use real-time transaction analysis to detect fraud patterns and stop unauthorized transactions.

• Algorithmic Trading: Real-time data processing enables financial firms to adjust trading algorithms instantly based on market changes.

• Customer Interaction: AI-powered chatbots and customer service agents are able to offer real-time support to clients, improving the customer experience.

Data Security and Compliance in Financial Data Handling

In finance, data is not just an asset – it is also a liability. Financial institutions need to adhere to strict data security and compliance regulations to protect sensitive customer information and meet legal requirements.

Compliance with Regulations

Financial institutions operate in a heavily regulated environment, where maintaining compliance is crucial. Regulations like GDPR (General Data Protection Regulation), FINRA (Financial Industry Regulatory Authority), and the SEC (Securities and Exchange Commission) set strict guidelines for how financial data should be handled, stored, and protected.

• GDPR: This European regulation imposes heavy fines on organizations that mishandle personal data. Financial institutions must ensure that they collect, store, and process customer data in compliance with GDPR principles, such as obtaining explicit consent and providing data access rights to users.

• FINRA/SEC Regulations: These U.S.-based regulatory bodies require firms to retain records of transactions and communications, ensure that data is protected from unauthorized access, and report suspicious activities promptly. Financial firms must implement stringent data governance frameworks to comply with these regulations.

Data Security in Financial Institutions

With the massive amount of sensitive data stored in financial systems, protecting this data from cyberattacks, breaches, and unauthorized access is of paramount importance. Financial institutions are leveraging a combination of encryption, multi-factor authentication (MFA), and access control policies to ensure the security of their systems.

• Encryption: Financial data, both at rest and in transit, is encrypted to prevent interception by malicious actors.

• MFA: Multi-factor authentication ensures that even if an attacker gains access to a password, they still cannot access the data without a second form of authentication (such as a token or biometric verification).

• Data Masking: Sensitive customer data, such as credit card numbers or Social Security numbers, is often "masked" in non-production environments to prevent accidental exposure during testing or development.

Cloud Platforms in Financial Data Handling: AWS, GCP, and Azure

Cloud platforms such as Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure have become the backbone for modern financial data management. These platforms offer scalable infrastructure, advanced analytics tools, and machine learning services that are essential for financial institutions to stay competitive.

Benefits of Cloud Platforms in Finance

• Scalability: Cloud platforms provide virtually unlimited storage and computing power, allowing financial institutions to scale operations efficiently.

• Security and Compliance: Major cloud providers offer industry-specific compliance certifications (such as SOC 2 or ISO 27001) and implement strong security features, including encryption and access control, to meet financial regulatory standards.

• Advanced Analytics and Machine Learning: Cloud platforms provide access to a range of tools for big data processing, AI model development, and real-time analytics. For instance, AWS provides services like Amazon SageMaker for machine learning, while Google Cloud’s BigQuery offers fast data analytics.

Use Cases of Cloud in Finance:

• Risk Analytics: Financial firms use cloud platforms to run complex risk simulations at scale, allowing them to identify potential vulnerabilities in their portfolios and strategies.

• Fraud Detection and Prevention: Cloud-based AI models can analyze billions of transactions in real time, flagging suspicious activities with greater accuracy than traditional systems.

• Customer Service Automation: Using cloud-based AI and chatbots, financial institutions can offer 24/7 customer service, streamlining support while reducing operational costs.

In the financial industry, leveraging the right data infrastructure is key to gaining a competitive edge. By effectively managing data using data lakes, data warehouses, and advanced cloud platforms, financial institutions can enhance their decision-making capabilities, improve security and compliance, and deliver a better experience to customers.

As the industry continues to embrace real-time data ingestion, advanced analytics, and AI, those who master the art of data management will be the leaders of tomorrow’s financial ecosystem.

Chapter 5: The Science Behind the Models – ML, NLP, and Predictive Analytics

Artificial Intelligence (AI) in finance is not magic – it’s applied science. Behind every real-time fraud alert, automated investment strategy, or smart credit score is a complex stack of algorithms and data pipelines.

To make AI work in financial environments where accuracy, explainability, and risk tolerance are non-negotiable, institutions rely on a blend of machine learning (ML), natural language processing (NLP), and predictive analytics.

In this section, we’ll unpack the foundational AI methods that power today’s most critical financial systems, and how these models are reshaping decision-making across the value chain.

Time-Series Forecasting: The Engine of Financial Prediction

Time-series forecasting is the cornerstone of financial modeling. Unlike typical supervised learning where inputs are independent, time-series models take into account temporal dependencies – the past influencing the future – which is especially important in domains like stock prices, interest rates, and credit defaults.

Core Applications in Finance:

• Asset Price Prediction: Hedge funds and asset managers forecast equity, FX, and commodity prices using techniques ranging from ARIMA and exponential smoothing to deep learning-based models like LSTMs (Long Short-Term Memory) or Temporal Convolutional Networks (TCNs).

• Liquidity Forecasting: Treasury departments forecast cash flow and liquidity needs across accounts and geographies to meet regulatory buffers and prevent shortfalls.

• Credit Risk Monitoring: Time-series models help anticipate changes in borrower behavior or macroeconomic indicators that impact default probabilities.

Technical Insights:

• Models Used: ARIMA, Prophet (developed by Meta), LSTM, XGBoost on rolling features.

• Challenges: High noise-to-signal ratio in markets, non-stationarity, and the risk of overfitting to past data.

• Best Practices: Combining feature engineering with domain-specific constraints (for example, market open/close calendars, economic events) significantly improves forecast reliability.

Risk Modeling: Quantifying Uncertainty with Machine Learning

Risk modeling is fundamental in finance, whether you're managing market risk, credit risk, or operational risk. Traditionally built with logistic regression and rule-based systems, today’s models are becoming far more nuanced through ML.

Machine Learning in Risk:

• Credit Risk: ML models ingest not just FICO scores and payment history, but also alternative data like cash flow, mobile phone usage, and behavioral patterns to score borrowers – especially useful in emerging markets or for thin-file customers.

• Market Risk (VaR, CVaR): ML techniques simulate potential portfolio losses under different market scenarios, accounting for complex correlations across assets.

• Operational Risk: Using internal logs and incident reports, anomaly detection algorithms can flag early indicators of system failures or fraud.

Technical Highlights:

• Popular Models: Gradient Boosting Machines (GBM), Random Forests, Support Vector Machines (SVM), and Neural Networks.

• Interpretability: Risk models must be explainable to pass regulatory scrutiny. Tools like SHAP values or LIME help demystify black-box models by showing the impact of individual features on predictions.

• Example: A bank may use XGBoost to predict credit card default, with SHAP showing that recent missed payments and high utilization ratios were the key drivers behind the model’s output.

Natural Language Processing (NLP): Unlocking Textual Data

Financial institutions sit on mountains of unstructured textual data — earnings call transcripts, analyst reports, regulatory filings, news, and customer communications. NLP allows them to extract meaningful insights from this data at scale.

Use Cases in Finance:

• Document Review and Contract Analysis: NLP models scan thousands of legal agreements or credit contracts to flag risk clauses, expirations, or inconsistencies (for example, JPMorgan’s COiN platform).

• Sentiment Analysis: Hedge funds use NLP to analyze news and social media sentiment to anticipate market movements.

• Regulatory Compliance: Automated systems parse SEC filings, GDPR policies, and internal communications to ensure compliance or detect violations.

• Customer Service Chatbots: NLP powers real-time customer engagement, automatically resolving queries and routing issues to the right departments.

Technologies:

• Traditional Methods: Named Entity Recognition (NER), Bag-of-Words, TF-IDF, Latent Dirichlet Allocation (LDA).

• Modern Approaches: Transformer models (like BERT, RoBERTa, or domain-specific variants such as FinBERT) trained on financial texts to achieve better context understanding.

• Document Intelligence: With models like GPT-4 or Claude, banks can now extract and summarize key risks, opportunities, or inconsistencies from dense reports.

Fraud Detection: Using Anomaly Detection and Unsupervised Learning

Fraud detection is one of the highest ROI use cases for AI in finance. The challenge lies in identifying non-obvious, evolving fraudulent patterns buried in billions of transactions – often without labeled data.

Why ML Outperforms Rule-Based Systems:

• Traditional systems rely on static rules like “flag any transaction over $5,000 abroad.” But fraudsters quickly adapt.

• Machine learning systems, particularly those using unsupervised or semi-supervised techniques, learn what “normal” looks like for each user and flag outliers in real-time.

Models and Approaches:

• Unsupervised Learning: Clustering (for example, DBSCAN), Autoencoders, and Isolation Forests are used to detect anomalies without needing labeled fraud data.

• Semi-Supervised Learning: Train on a small labeled dataset with millions of unlabeled records.

• Behavioral Biometrics: ML models monitor how users type, swipe, or move the mouse to detect suspicious behavior – often used in mobile banking apps.

Example:

A neobank like Revolut may apply autoencoder-based models on real-time transaction data. If a user who typically shops in Amsterdam suddenly makes 5 high-value transactions from São Paulo using a new device, the system flags and freezes the account for verification – all within milliseconds.

Behind every AI solution in finance is a combination of mathematical modeling, data engineering, and domain expertise. Whether it’s a hedge fund predicting earnings, a bank screening loans, or an insurance firm processing claims, these tools – time-series forecasting, ML-based risk scoring, NLP-driven document analysis, and anomaly detection – are the technical foundation of financial AI. Understanding them is not optional for executives anymore – it’s the difference between leading innovation or being disrupted by it.

Chapter 6: Training the Workforce – Upskilling Executives, Technical, and Non-Technical Teams in FinTech

AI transformation in finance is both a technological shift and an organizational one. Success doesn’t depend solely on algorithms or data pipelines, but on people: the ones who design, deploy, fund, govern, and use AI.

And if there's one hard truth in AI transformation, it is this: Innovation starts at the top.

Whether you are running a regional bank, a global asset manager, or a fintech startup, your leaders must be AI-literate. Not necessarily technically fluent in code – but strategically fluent in AI’s business value, risks, and implementation realities.

AI Literacy for Leadership: A Strategic Imperative

The idea that AI is a luxury – or something to “consider later” – is a dangerous misconception. In the current financial landscape, AI is a necessity. And if decision-makers don’t understand it, they can’t lead it.

Executives are the ones who sign off on technology budgets, approve digital initiatives, and set strategic priorities. It doesn't matter how innovative your engineers are. If your leadership doesn’t “get” AI, the innovation dies on the boardroom table.

Common Executive Blind Spots:

• Confusing automation with true AI (for example, rules-based tools vs. learning systems)

• Underestimating the cost and complexity of model deployment

• Failing to understand data infrastructure dependencies

• Viewing AI as a “tech problem” instead of a business enabler

• Ignoring governance risks or regulatory exposure

Here are some key topics in executive AI training:

• Understanding ML, NLP, and GenAI at a strategic level

• Interpreting AI project KPIs and business ROI

• Governance and model risk management

• Ethical and regulatory frameworks (EU AI Act, GDPR, SEC AI enforcement)

• Building cross-functional AI innovation teams

"You’re not going to lose your job to an AI, but you’re going to lose your job to someone who uses AI."
— Jensen Huang

This is not hyperbole. It's already happening. In a 2024 survey by PwC, 72% of financial services CEOs admitted they lacked a clear understanding of how AI delivers ROI in their own organizations. Meanwhile, 60% of digital transformation failures in banking were attributed to “leadership misalignment”, not technical challenges.

The Cost of Inaction:

• Slower go-to-market for AI-based products

• Missed competitive advantages (for example, predictive credit scoring, customer retention models)

• Increased risk of non-compliance due to lack of AI governance

• Talent attrition – top AI engineers don’t stay where innovation is blocked

To address this, top-tier financial institutions are increasingly mandating structured AI education programs for senior leaders, including CEOs, CTOs, COOs, and board members. This isn't just optional professional development – it's often required to ensure alignment on AI strategy, ethical use, and ROI measurement.

Why Mandating AI Education is Becoming Standard

The push for mandatory AI training stems from several factors:

1. Strategic Imperative

A 2024 PwC survey cited in various reports notes that 72% of financial services CEOs lack a clear understanding of AI's ROI, contributing to 60% of digital transformation failures due to leadership misalignment. Mandated programs help bridge this by providing strategic fluency in machine learning (ML), natural language processing (NLP), generative AI, and regulatory frameworks like the EU AI Act or GDPR.

2. Risk Mitigation

With AI introducing new risks (for example, bias in models, data privacy breaches), boards and executives need education to oversee governance. For instance, the Global Financial Stability Board warned in 2024 that inconsistent AI standards could pose systemic risks.

3. Competitive Edge and Talent Retention

Institutions that invest in executive education see faster AI adoption, better talent attraction, and reduced attrition. Training costs (for example, $5,000 per person annually) are often offset by savings from avoiding missteps, as outlined in the handbook.

4. Regulatory and Market Pressures

Bodies like the FDIC and OCC have released training resources (for example, FDIC videos on cybersecurity for bank directors), signaling expectations for AI literacy. Conferences like the 2024 FSOC AI & Financial Stability event and Opal Group's Compliance in the Age of AI 2025 emphasize executive involvement.

These programs typically cover AI fundamentals, use cases in finance (for example, predictive analytics), ethical considerations, and hands-on tools like ChatGPT or custom platforms. Formats range from in-house workshops and reverse mentorships to external certifications and business school courses.

Institutions and Executives Mandating AI Education

While adoption varies by region and institution size (stronger in the US and Asia, as you may be able to tell), several top-tier players are leading with mandated or structured programs. Let’s look at some key examples drawn from recent developments as of July 2025:

1. Bank of America: The bank has adopted a top-down approach to AI education, mandating briefings for senior leadership on generative AI's potential and risks starting around 2023. This includes required sessions for executives to understand AI integration in retail, small business, and wealth management. Hari Gopalkrishnan, CIO and Head of Retail, Small Business, and Wealth Technology, leads this initiative, ensuring C-suite alignment to drive efficient operations and mitigate risks. This reflects a broader trend where banks prioritize internal AI tools for employee training, extending to executives.

2. Morgan Stanley: As a pioneer in AI deployment (for example, their COiN platform mentioned above), Morgan Stanley integrates mandatory AI training into tool rollouts for wealth management teams, including executives. Tools like the Morgan Stanley Assistant (launched September 2023, powered by OpenAI's GPT-4) and Morgan Stanley Debrief (June rollout) require user training embedded in the experience. Koren Picariello, Managing Director and Head of Wealth Management Generative AI, oversees this, emphasizing intuitive learning for financial advisors and support staff – though it extends to leadership for strategic oversight. This approach ensures executives are fluent in AI to support firm-wide adoption.

3. Community Financial Institutions (CFIs) via Eltropy: Credit unions and community banks are mandating AI certification through Eltropy's program, launched post-EMERGE 2025 conference where over 130 professionals earned the Eltropy AI Practitioner Certificate. This self-paced, on-demand certification is required for employees across functions, including executives, covering foundational AI, Agentic AI, compliant usage in regulated environments, and hands-on bot-building with technologies like LLMs and prompt engineering. While not naming specific executives, it's tailored for CFI leaders to build and deploy AI immediately, addressing the handbook's call for upskilling in smaller institutions.

4. General Banking Boards (for example, via BankDirector Guidance): Many US banks mandate director education and onboarding focused on AI skills for board members to oversee implementation effectively. This includes reboarding programs to enhance technology expertise, with boards establishing governance committees and designating AI overseers. For example, boards are encouraged to support capital for AI infrastructure while receiving regular updates, ensuring members are trained to guide ethical integration and competitive strategies.

5. Hedge Funds and Larger Institutions: A 2024 AIMA report on hedge funds shows that nearly half of larger managers (for example, those managing significant AUM) mandate Gen AI training for teams, including executives, though overall adoption is at 10% industry-wide. Firms like Citadel, Bridgewater Associates, and Renaissance Technologies (highlighted in Senate investigations) are creating multidisciplinary AI teams, implying required upskilling for quants and leaders. Bridgewater's CEO, Nir Bar Dea, has publicly discussed AI's role in altering hedge fund landscapes, suggesting internal education mandates.

6. Broader Trends Involving CEOs and Boards: Across sectors, boards and CEOs are forming joint AI vision task forces that mandate quarterly meetings and ethical scorecards, often including reverse mentorship programs where board members pair with AI specialists for hands-on learning. Business schools are incorporating AI case studies into board training, as noted in WSJ reports, to address a 20% tech expertise gap per PwC. Advisory firms like RSM US recommend CEOs and boards seek external education for AI vision-building, with 67% of organizations needing outside help.

These examples illustrate a shift toward mandatory AI literacy at the highest levels, aligning with our emphasis on transforming executives into innovation champions. Institutions like Bank of America and Morgan Stanley exemplify how this combats hesitation, fostering a culture where AI drives measurable value.

Training Technical Teams in FinTech

While AI literacy for leadership is essential, innovation doesn’t happen from the boardroom alone. It must be embedded across technical teams – engineers, analysts, data scientists, and product professionals – who build and maintain the infrastructure for change.

But here’s the critical point: you cannot innovate with an exhausted, overburdened, and undertrained workforce.

Many companies today are asking their software engineers to become AI engineers overnight. They're assigning responsibilities for data science, MLOps, predictive modeling, or chatbot design to backend developers who lack the training to handle data pipelines, model deployment, or even fundamental AI architecture. This isn't just inefficient – it's a recipe for failure.

Why Upskilling Pays Off

Let’s look at this through the lens of hard numbers.

A company with a technical team of 100 software engineers, data scientists, or IT professionals will, on average, lose 13 team members per year. For every engineer who leaves, the cost of replacement – including hiring, onboarding, training, lost productivity, and project disruption – averages $83,000. That means the company loses around $1.08 million per year due to attrition alone.

And this figure only reflects direct costs. It doesn’t include lost time on strategic initiatives, intellectual capital, or the hidden tax of slower innovation. These losses compound over time – especially when the market is rapidly adopting AI and you're left with gaps in capability.

Now compare that with the cost of strategic upskilling.

If you invest in targeted AI and data training at a rate of $5,000 per person per year, your total investment for 100 engineers is $500,000 per year. That’s less than half the cost of attrition.

But the ROI is even bigger when you account for what you gain:

• Stronger employee retention (engineers are more likely to stay when growing their skill set)

• Faster delivery of AI-powered features, internal tools, and customer experiences

• Reduced need to hire external consultants or chase niche AI talent in a hyper-competitive market

• Avoiding expensive failures caused by technical debt or improperly built models

When engineers are trained in areas like machine learning, LLM integration, NLP, MLOps, and data pipelines, they become innovation enablers rather than just code executors.

Hidden Cost of Overburdening Engineers

What many executives don’t realize is that undertrained engineers – especially when asked to build high-risk AI systems – can expose the company to massive business risk. They may build flawed recommendation systems, opaque risk models, or chatbot interactions that spiral into compliance disasters.

Modern AI systems require more than good coding skills. They also require:

• Deep understanding of how to clean, structure, and prepare data

• Familiarity with supervised vs. unsupervised learning

• Knowledge of transformer models, fine-tuning, vector search, embeddings

• Awareness of AI ethics, explainability, and regulatory frameworks

These skills are not taught in traditional software engineering programs, nor are they something engineers can "pick up on the job" during sprints. Asking your developers to do everything – from backend infrastructure to building black-box models – is not only unfair, it’s strategically reckless.

Upskilling Is Not a Cost — It’s a Hedge Against Brain Drain

Here’s the basic math again:

• Cost of attrition per year (100 engineers, 13 lost): $1,079,000

• Cost of upskilling per year (100 engineers, $5K each): $500,000

• Net savings from upskilling: $579,000 annually

And this is before counting the additional business value from faster launches, higher employee morale, and innovation that drives new revenue streams.

Investing in upskilling not only saves you money – it future-proofs your talent pipeline and makes your team more self-sufficient. Engineers who stay and grow are more likely to build products that push your business forward.

Motivation Through Growth

One of the most overlooked retention strategies in tech is personal and professional development. Talented engineers want to work at companies where they grow. When organizations ignore this, they create frustration, stagnation, and ultimately attrition.

On the other hand, those who invest in upskilling create a sense of purpose and momentum. Upskilled engineers are more confident, more collaborative, and more likely to take initiative in applying AI to business problems.

Training isn't a perk – it's a competitive edge.

Training Non-Technical Professionals: Empowering the 95% with AI Fluency

In the conversation around AI transformation, technical talent gets much of the attention – and rightly so. But the reality is this: 95% of the workforce in most organizations is not technical. And yet, 95% of employees are now asking for training in generative AI, according to a 2024 global workplace survey by edX and The Harris Poll.

This signals a shift in awareness: non-technical professionals understand that generative AI isn’t just a tool for developers – it’s a work enhancer, a productivity multiplier, and a competitive necessity.

From Fear to Fluency: Why Non-Tech Training Matters

The fear narrative around AI – that it will take away jobs – is real and palpable in many organizations. But the more strategic view is this:

Don’t fire your workforce. Train them.

Rather than replacing administrative staff, compliance officers, relationship managers, operations teams, and analysts, leading financial organizations are upskilling their existing talent to work with AI, not against it.

Training non-technical team members in generative AI offers two major business advantages:

1. Productivity gains: Teams can automate repetitive, low-value tasks and focus more on decision-making and strategy.

2. Talent retention: Employees feel more secure and valued when their employers invest in their future.

Use Cases: Where Non-Tech Teams in Finance Can Gain from AI Training

Non-technical employees in banking, asset management, insurance, and fintech can immediately apply generative AI tools across their workflows. Here’s how:

1. Compliance & Legal Teams

• Use ChatGPT or Claude to summarize regulatory documents, contracts, and internal audit reports.

• Use Phoenix to draft standard policies and regulatory templates, saving hours of manual editing.

• Extract key clauses from loan agreements or KYC policies.

• Draft internal memos or SAR summaries 2–3x faster.

2. Finance, Accounting, and Operations

• Automate spreadsheet generation and financial modeling using Microsoft Copilot in Excel.

• Reconcile data from multiple sources and generate summary reports.

• Draft and revise standard Jira tickets or issue documentation using Phoenix, bridging business and IT communication.

3. Sales, Relationship Management, and Customer Service

• Use generative chat tools to personalize client interactions.

• Draft follow-up emails, presentations, and pitch summaries.

• Summarize meeting transcripts and extract actionable items.

4. Marketing and Communications

• Use AI to generate segmented content for different client audiences.

• Produce A/B tested campaign text, product updates, and social posts.

• Translate campaigns quickly for global markets.

5. Risk & Audit

• Summarize findings from large datasets or transaction logs.

• Generate first-draft risk assessments and credit memos.

• Highlight inconsistencies or anomalies with contextual explanation.

The Cost of Not Training: A Missed Opportunity

Non-technical employees touch every part of your organization – operations, client relations, document handling, and decision support. If they are not AI-enabled, your business is flying with one wing.

Training these employees doesn't mean turning them into engineers. It means:

• Teaching them how to interact effectively with AI

• Helping them become critical evaluators of AI output

• Guiding them to avoid over-reliance or misuse of AI tools

This form of AI literacy is the new digital literacy – essential for everyone, not just technologists.

Chapter 7: AI for Executives, AI Education & Enablement in Finance – Workshops, Tools, Services, and Training Resources

The most innovative financial institutions no longer see AI training as a "nice-to-have." In an increasingly algorithmic economy, where generative AI tools are reshaping everything from compliance to capital allocation, AI education is an investment in strategic resilience.

This section offers a clear, credible breakdown of how to get your teams – executive and operational – up to speed through trusted workshops, tools, agencies, and courses. It emphasizes the value of enabling internal transformation instead of relying solely on outside hires.

AI Certifications for Banking Professionals

Several industry and educational organizations offer certification programs specifically designed for finance professionals:

1. Generative AI In Finance and Banking Certification: This program teaches applications of generative AI models, including generative adversarial networks (GANs) and transformers for predicting market trends, automating financial tasks, and enhancing customer experiences. You can learn more about the cert here.

2. Certificate in Digital & AI Evolution in Banking: This certification helps professionals understand the digital transformation in banking, including regulatory considerations and the risks and benefits of technology adoption. You can learn more about the cert here.

3. Machine Learning for Investment Professionals: Offered by the CFA Institute, this program focuses on machine learning applications specifically for investment management and analysis. You can learn more about the Investment Management with Python and Machine Learning specialization here, and the CFA Institute Machine Learning course here.

Columbia Business School's AI for Business & Finance Certificate Program is particularly noteworthy, as it "has been designed for professionals in the business and finance world who need to learn AI but don't really have a technical background". This eight-week course covers AI fundamentals, Python programming for finance, predictive analytics, and generative AI business applications.

Conclusion

In an era where artificial intelligence is reshaping the financial landscape, executives and teams need to recognize that adapting to AI is not just a strategic advantage – it's a survival imperative. Just as we've successfully navigated previous technological revolutions, from the internet and cloud computing to blockchain and big data, AI presents an opportunity to democratize access to cutting-edge tools, empowering a broader range of professionals to innovate in ways that were once unimaginable.

This inclusivity has already sparked breakthroughs in predictive analytics, risk management, and personalized services, allowing even smaller institutions to compete on a global scale. That said, AI's integration into finance is far from novel. Leading institutions have deployed these technologies for years, embedding them into core operations like fraud detection and algorithmic trading.

Yet, for newcomers or those refreshing their approach, the relevance remains profound. Ongoing updates and advancements – such as enhanced natural language processing models and real-time data ingestion capabilities – continually amplify the potential for investment managers, AI specialists, and broader teams, unlocking efficiencies and insights that elevate professional capabilities to new heights.

To harness this potential and maintain a competitive edge, continuous upskilling is essential. Executives and teams alike should commit to updating their knowledge base through targeted education programs, workshops, and resources, ensuring they stay ahead of the curve.

Ultimately, AI can be a force for profound good. At LunarTech, we don't foresee it leading humanity to doom – instead, in a world facing complex challenges like economic volatility and climate risks, AI stands as a powerful ally, one that could very well guide us toward solutions and a brighter future. By embracing it thoughtfully, the financial sector can lead this transformation, fostering innovation that benefits all.

Newsletters to Follow for FinTech

Our Newsletter

LUNARTECH Newsletter - https://lunartech.substack.com/

US Personal Finance & Investment Newsletters

• Money Stuff (Matt Levine, Bloomberg): Witty, in-depth takes on Wall Street and finance.

• TKer (Sam Ro): Stock market insights and long-term investment themes.

• Jill on Money (Jill Schlesinger): Financial news and expert advice, weekly.

• Behavior Gap (Carl Richards): Simple sketches and insights on money and decision-making.

• The Minority Mindset / Market Briefs (Jaspreet Singh): Daily, concise financial news and wealth-building tips.

• Exec Sum (Litquidity): Quick, reliable summaries of major finance news.

Baltic & Regional Newsletters

• Fintech News Baltic: News and trends in Baltic fintech, startups, and digital finance.

• Linas Beliūnas – FinTech Digest (LinkedIn): Personal insights on fintech, AI, and digital assets from a leading Lithuanian expert.

• Change Ventures Weekly: Baltic startup and VC news, funding rounds, and hiring.

• CFO Club Newsletter: Modern finance newsletter for tech sector CFOs and leaders-trends, tips, and innovation.

LunarTech AI for Executives

For leaders and frontline professionals who feel the pressure to “get AI” but don’t speak code, this 1- to 3-day program delivers exactly what you need: no fluff, no jargon. In clear language, we unpack how generative AI, large-language models, and regulatory frameworks such as the EU AI Act are reshaping compliance, risk, and client service.

Next, we roll up our sleeves. You’ll practice with ChatGPT, Phoenix, Gemini, and other curated tools to summarize 200-page reports in minutes, flag hidden risks, and automate repetitive workflows. Expect live demos, breakout labs, and case studies drawn straight from banking, asset management, and insurance.

By the final session you’ll have a road-ready playbook for piloting AI safely – from data-governance checklists to ROI metrics your CFO will love. Graduates leave with a certificate, a toolkit of prompts, and the confidence to champion AI initiatives inside their own departments.

• Format: Online or on-site, 1–3 days

• Cost: $997 per participant

Apply Here: https://lunartech.ai/programs/ai-for-executives

LunarTech Academy

Our Academy is the always-on learning hub that keeps finance professionals current long after the headlines fade. Courses are modular and industry-specific, so a portfolio manager can master forecasting in Python while a relationship manager explores generative-AI productivity hacks – all under one roof.

Every track is written by practitioners who ship models in production, not theorists. Expect bite-size videos, step-by-step notebooks, and capstone projects pulled from real trading, risk, and compliance datasets. Learners can move at their own pace or join live cohorts for instructor feedback and peer discussion.

Managers love us for the built-in LMS integration, progress analytics, and team licensing that scales from five seats to five hundred. Whether you need to onboard new hires fast or reskill an entire division, the Academy delivers measurable, trackable outcomes.

• Format: Self-paced or instructor-led; team licenses available

• Cost: $49.97 – $199.97 per month

Apply Here: https://academy.lunartech.ai/

Other Resources

• Lens | LUNARTECH - https://lens.lunartech.ai/

• YouTube | LUNARTECH - https://www.youtube.com/@lunartech_ai

• Linkedin | LUNARTECH - https://www.linkedin.com/company/lunartechai/

• Substack | LUNARTECH - https://lunartech.substack.com/

You’ve been grinding through tutorials. You've built a portfolio site, maybe deployed a few projects on GitHub. And now you're trying to land a job or join a team.

Then the interviews start.

Suddenly, people ask:

• "Are you familiar with Agile?"

• "Have you worked in a Scrum environment?"

• "What’s your experience with sprints?"

Cue the imposter syndrome. Because no one teaches this stuff in JavaScript 101.

This guide is for you.

I’ll help make the Scrum process – a very common way developers work together – make actual sense. I’ll walk you through the basics, but also tell you what developers actually do, how standups feel when you're new, and what’s expected of you when you’re no longer coding in a vacuum.

Let’s break it down.

Here’s what we’ll cover:

• What Even Is Scrum?

• The Three Roles in Scrum (and Who Does What)

• The Scrum Rhythm: What a Sprint Actually Looks Like

• Who attends the Ceremonies:

• Standups: Where You Talk Like a Human, Not a Robot

• Sprint Planning

• What’s a User Story and Why Does It Sound Like a Children’s Book?

• What Counts as “Done”? Definition of Done and Why It’s Important

• Demos, Retros, and Saying the Hard Things

• Tools You Might Encounter

• If You’re Preparing for a Job, Here’s What You Can Do

• Final Thoughts

What Even Is Scrum?

Scrum is not a tool. It’s not a software. It’s not some elite thing only PMs care about.

It’s a lightweight framework that helps software teams build things incrementally, together, in short focused cycles called sprints.

Scrum is used by everyone from FAANG teams to indie dev shops because it helps:

• Keep teams aligned

• Deliver working software fast

• Course-correct often

• Spot problems early (before they go nuclear)

It’s the opposite of the old-school “build for a year and pray it works” model.

The Three Roles in Scrum (and Who Does What)

Scrum officially defines three roles. Here's what that means in practice:

1. Product Owner (PO)

Think: Vision-holder. They decide what the team builds and why. A product owner:

• Writes user stories (think of these as feature requests written from a user’s point of view)

• Prioritizes the work

• Clarifies what success looks like

• Says “yes” or “not yet” to features

2. Scrum Master (SM)

Think: Air-traffic controller meets therapist. They make sure the process works. The are master Facilitators, like between Dev and PO’s. A Scrum Master:

• Facilitates meetings

• Removes blockers (“Your AWS access is stuck? I’ll escalate it.”)

• Coaches the team on Scrum practices

• Doesn’t manage people – manages flow

3. Developers (YOU!)

Think: Builders. You write code, test it, ship it, fix it, and improve it. You also:

• Break down stories into tasks

• Pick up work from the team board (like Jira or Trello)

• Communicate progress

• Demo what you’ve built at the end of the sprint

You might also work with designers, testers, or DevOps folks – but within Scrum, you’re all “developers” building a product together.

The Scrum Rhythm: What a Sprint Actually Looks Like

Image Source: https://www.invensislearning.com/blog/what-are-scrum-ceremonies/

Understanding the Scrum Cycle

So, what does it actually look like when a team uses Scrum to build software?

Let’s walk through a full sprint – not just the buzzwords, but what really happens when a group of humans tries to plan, build, review, and improve together. Think of this as your backstage pass to the rhythm of modern teamwork.

📦 Step 1: Build and Refine the Product Backlog

Before any coding starts, the team needs to agree on what they might build – not just this week, but in the near future too.

That’s where the Product Backlog comes in. This is a big, running list of everything the product might need – features, bug fixes, improvements, ideas, and maybe a few wild dreams. It’s like the wishlist for the product, but more organized (ideally).

The Product Owner is responsible for maintaining and prioritizing this list. They decide what’s most important to work on based on customer needs, business goals, and feedback.

But the PO doesn’t do this in isolation. Enter the Backlog Refinement meeting.

In these sessions, the Scrum Team – that’s the PO, the Scrum Master (SM), and the Developers – come together to:

• Review the most important upcoming items

• Clarify any vague or confusing parts of each task

• Break big items down into smaller, buildable pieces called user stories

• Estimate effort (how much time or complexity is involved for each story)

This meeting makes sure the team isn’t caught off guard in the sprint – that they understand the work ahead and can actually start sprinting when the time comes.

🧭 Step 2: Sprint Planning – What Are We Building This Time?

Now that we’ve got a solid backlog, it’s time to pick what to build right now.

At the start of each sprint (which typically lasts 1 to 4 weeks), the team holds a Sprint Planning meeting. This meeting sets the stage for the entire sprint – it’s like the huddle before the big game.

In Sprint Planning, the team:

• Reviews the top items from the backlog

• Discusses what can realistically be completed based on their availability and capacity

• Chooses a handful of these stories to commit to

• Defines a Sprint Goal – a simple statement that captures the purpose of this sprint

For example, the Sprint Goal might be:
🎯 “Allow users to reset their passwords.”

Every user story chosen should contribute to that goal. The collection of these stories becomes the Sprint Backlog – basically, the to-do list for the sprint.

So when we say:

“The team selects an ordered list of user stories to comprise the Sprint Backlog for the next sprint, which will be achievable to satisfy the Sprint Goal...”

We’re really just saying:
👉 “Pick a realistic number of important tasks that, if completed, will help us hit our target for the sprint.”

Not too vague. Not too ambitious. Just achievable and focused.

☀️ Step 3: Daily Standups – Stay in Sync

Now the sprint is underway! But how does everyone stay aligned and avoid working in silos?

That’s where the Daily Standup comes in. Every day – usually in the morning – the team has a quick check-in (about 15 minutes) where each person answers three questions:

1. What did I do yesterday?

2. What am I working on today?

3. Is anything blocking me? (that is, am I stuck?)

Example:

“Yesterday I set up the login API integration. Today I’ll work on the UI validation. I’m blocked on getting access to the staging database — may need help.”

These standups keep the team in sync and surface blockers early so they can be addressed quickly. They’re not about micromanaging or showing off. They’re about visibility and support.

📉 What’s a Sprint Burndown Chart?

You might hear your team mention a “burndown chart.” No, this isn’t about things going down in flames (hopefully).

A Sprint Burndown Chart is a graph that shows how much work is left in the sprint – day by day.

• The y-axis is the amount of work remaining (often measured in story points or tasks)

• The x-axis is the number of days left in the sprint

The line should ideally trend downward as work gets completed – hence “burning down.” If it flattens out or slopes up, that’s a red flag that the team might be stuck, behind schedule, or not updating the board.

Think of it as a visual heartbeat of the sprint. You can learn more via a practical example in this video.

🖥️ Step 4: Sprint Review – Show What You’ve Built

At the end of the sprint, the team holds a Sprint Review (also called a “demo”). This is where you show what was actually built during the sprint.

• The Developers demo working features – live, not just screenshots

• The Product Owner reviews whether the Sprint Goal was achieved

• Stakeholders may ask questions, give feedback, or suggest tweaks

This meeting isn’t just for show – it’s a feedback loop. It helps the team validate that what they built is useful, usable, and meets expectations. If changes are needed, those get added to the backlog for future sprints.

🔍 Step 5: Sprint Retrospective – Look Back to Move Forward

Once the review is done, the team shifts focus from what they built to how they worked together.

Enter the Sprint Retrospective – a meeting to reflect on the process, not the product.

The team discusses:

• ✅ What went well

• ❌ What didn’t go so well

• 🔁 What could be improved next time

This isn’t about pointing fingers. It’s about learning, adapting, and continuously improving how the team collaborates.

The Scrum Master often facilitates this meeting and helps turn feedback into action items for the next sprint. For example:

“We underestimated testing time. Next sprint, let’s budget for QA earlier.”

The best teams take retros seriously. Why? Because even if your code is perfect, your process needs tuning too – and small process changes often lead to big gains.

♻️ Scrum Is a Loop

Here’s the rhythm:

1. Plan the sprint

2. Check in daily

3. Build and demo the product

4. Reflect and improve

Then do it all over again – with slightly better coordination and slightly more trust each time.

It’s not about being fast. It’s about being intentional, consistent, and collaborative.

Example Sprint

Let’s say, for example, that your team does 4-week sprints. (Keep in mind that Sprints can differ by team, nature of product, release cycles, and so on.)

Here’s the rough beat:

Scrum works in loops. Every 2-4 weeks (depending on your cadence and sprint cycle), your team should have working, demo-able software to show for it – even if it’s small.

And no, it’s not about “speed.” It’s about consistency, communication, and collaboration.

Who attends the Ceremonies:

Now lets dive deeper and understand practically how each of these ceremonies work:

Standups: Where You Talk Like a Human, Not a Robot

So how does the team actually stay connected day to day? That’s where standups come in.

Every morning, your team meets briefly – usually on Zoom or in a circle – and you answer 3 questions:

1. What did I work on yesterday?

2. What will I work on today?

3. What’s blocking me? Any impediments?

Example:

"Yesterday I cleaned up the signup validation logic. Today I’m working on the email verification flow. I’m stuck on SendGrid config – might need help setting up credentials."

It’s not about impressing anyone. It’s about keeping everyone in sync. Some days you’ll say, “I spent the whole day debugging a CSS bug that turned out to be a semicolon.” That’s okay.

How does it work?

The Scrum Master gathers everyone in a huddle room, the PO and Dev Team included, and opens the the Standup. They are the facilitator of the ceremony. Everyone gets a chance to answer the 3 questions above (usually about 2-5 minutes each). It’s not a full report – it’s quick. When one person is done, they pass it on to someone else.

This ensures there is team cohesion and transparency.

Here is a video example of a standup.

Sprint Planning

The goal of the planning meeting is to answer the questions “What are we going to work on, and how are we going to do it?” It is critical for the team to have a shared goal and a shared commitment to this goal before beginning this ceremony.

Participants should:

• Measure growth

• Sync with the Scrum Master

• Sync with the Product Owner

Sprint planning happens just before the sprint starts, and usually lasts for an hour or two. In this meeting, the team goes over a collection of user stories and discuss, plan, measure, and prioritize. This is where they decide what is going to be in scope for their upcoming sprint cycle.

The Product Owner will have a prioritized view of things in the backlog. They work with the team on each object or customer experience. Together, as a group they go through and make calculations, deciding to what they can commit.

What’s a User Story and Why Does It Sound Like a Children’s Book?

So you might be wondering: how do you know what to work on? What to build? So much work, so little time? Thats where user stories come in.

In Scrum, teams don’t just write vague tasks like “code the login.” Instead, they write user stories – short, human-centered feature descriptions that describe what the user needs, why they need it, and what success looks like.

Here’s an example:

As a user, I want to be able to reset my password, so I can access my account if I forget it.

User stories are the scaffolding of teamwork. They’re written with empathy, not just efficiency. And each one comes with acceptance criteria – a checklist that clarifies what “done” actually means:

• A “Forgot Password” link is visible

• Clicking it shows a form

• An email gets sent with a reset link

Once a story is agreed upon, developers break it down into tasks, like “build form,” “hook into backend,” or “handle email validation.” It’s collaborative, not prescriptive. And user stories have priority so you know what’s the most important and what’s the least.

A helpful rule of thumb many teams use is the Gherkin-style "Given–When–Then" format:

• Given some initial context

• When an event occurs

• Then a specific outcome should happen

This ensures that everyone – devs, testers, and product owners – shares the same understanding of behavior and expectations.

Here is a great video example thats outlines how to draft effective and powerful user stories.

What Counts as “Done”? Definition of Done and Why It’s Important

Now you might be wondering – how do I know when a task is done and can be closed out?

The Definition of Done is a type of documentation in the form of a team agreement. The Definition of Done identifies the conditions that need to be achieved in order for the product to be considered done (as in potentially shippable).

This is how we know that we "did the thing right". Meaning, we built the correct level of quality into the product. The Definition of Done is not the same as the acceptance criteria, which are written by the product owner to help us know we did the "right thing".

Every team has a Definition of Done – it’s not just “I pushed code.” It could mean:

• Code is written

• Reviewed by a peer

• Merged into main

• Tested on staging

• Possibly deployed

This clarity keeps teams honest and accountable. No “it works on my machine” energy here. The DoD sets a quality bar. It prevents ambiguity, rework, and “it works on my machine” moments. When every card on the board passes the same finish line, teams move faster – and trust each other more.

Everyone should know what done is in a team. Either its Done as per DoD standards or its not.

Here is a beautiful video highlighting the impotence of DoD.

Demos, Retros, and Saying the Hard Things

Once you’ve built the product, then comes demos (showcasing your work) and retros (analysis as a team on what when well and what areas to improve on).

In the retro, everyone’s encouraged to speak up:

• What went well?

• What didn’t?

• What should we try next time?

Example:

“We missed a lot of stories because we didn’t account for testing time. Maybe we buffer next sprint with fewer tasks.”

The goal is not to blame – it’s to improve. Over time, this feedback loop becomes gold. The Scrum Master usually facilitates, collects feedback (via tools like Parabol, Miro, or sticky notes), and helps turn insights into actionable experiments for the next sprint.

Over time, retros become the heartbeat of team evolution.

Here is a video highlighting the importance of a Retro and Sprint Review.

🧠 Why Retrospection Matters More Than You Think

The Sprint Retrospective is more than just another meeting. It’s a mirror for your team – a safe, structured space to pause, reflect, and improve together.

You discuss:

✅ what went well

❌ what did not go well

🔁 what could we do better next time

Great teams don't just deliver great software, they continually deliver better ways of working.

This is why many experienced Scrum practitioners consider the retro to be the most important event in Scrum. Code is deployed once, but process improvements grow exponentially, sprint after sprint.

Tools You Might Encounter

Scrum doesn’t require software, but real-world teams use a variety of tools:

• Jira – Tracks sprints, issues, velocity

• Trello – Simple board, good for small teams

• Slack – Where standups often happen if async

• Notion / Confluence – Docs, retros, notes

• GitHub Projects – Lightweight planning for devs

Don’t worry if you’re not fluent in these yet. They’re tools – you’ll learn them on the job.

If You’re Preparing for a Job, Here’s What You Can Do

• ✍️ Practice writing user stories from your side projects

• 🧪 Run a mini-sprint: Plan your weekend project, set goals, and “review” it at the end

• 🤝 Contribute to an open-source project that uses Scrum or Agile workflows

• 🧾 Write about what you learned – maybe as a tutorial (hint hint)

Final Thoughts

So to recap, Scrum is a simple yet powerful way for teams to work together, stay organized, and deliver results quickly. It runs in short cycles called sprints, where the team plans what to do, checks in daily, shows their progress at the end, and reflects on how to improve.

The four key ceremonies – Sprint Planning, Daily Scrum, Sprint Review, and Sprint Retrospective – help keep everyone aligned and focused. With clear roles and regular feedback, Scrum makes it easier to handle changes, solve problems early, and continuously get better as a team.

But scrum isn’t a magic spell. It’s just a way for humans to build complex things – together – without falling apart.

You don’t need to be a Scrum Master. You don’t need a certification. But if you understand how sprints work, what’s expected of you, and how to show up to meetings with clarity and candor, you’re 10 steps ahead of most.

Scrum helps teams talk, plan, build, and learn. And now? You can too.

If you liked this, please do share. You never know who it might help out.

Until then…keep learning, unlearning, and relearning!!!

But large language models (LLMs) don’t actually have memory on their own. They don’t remember conversations unless that information is given to them again.

So, how do they seem to recall things?

The answer lies in something called a vector store – and that’s what you’ll learn about in this article.

Table of Contents

• What Is a Vector Store?

• How Embeddings Work

• Why Vector Stores Are Crucial for Memory

• Popular Vector Stores

  • FAISS (Facebook AI Similarity Search)

  • Pinecone

• Making AI Seem Smart with Retrieval-Augmented Generation

• The Limits of Vector-Based Memory

• Conclusion

What Is a Vector Store?

A vector store is a special type of database. Instead of storing text or numbers like a regular database, it stores vectors.

A vector is a list of numbers that represents the meaning of a piece of text. You get these vectors using a process called embedding.

The model takes a sentence and turns it into a high-dimensional point in space. In that space, similar meanings are close together.

For example, if I embed “I love sushi,” it might be close to “Sushi is my favourite food” in vector space. These embeddings help an AI agent find related thoughts even if the exact words differ.

How Embeddings Work

Let’s say a user tells an assistant:

“I live in Austin, Texas.”

The model turns this sentence into a vector:

[0.23, -0.41, 0.77, ..., 0.08]

This vector doesn’t mean much to us, but to the AI, it’s a way to capture the sentence’s meaning. That vector gets stored in a vector database, along with some extra info – maybe a timestamp or a note that it came from this user.

Later, if the user says:

“Book a flight to my hometown.”

The model turns this new sentence into a new vector. It then searches the vector database to find the most similar stored vectors.

The closest match might be “I live in Austin, Texas.” Now the AI knows what you probably meant by “my hometown.”

This ability to look up related past inputs based on meaning – not just matching keywords – is what gives LLMs a form of memory.

Why Vector Stores Are Crucial for Memory

LLMs process language using a context window. That’s the amount of text they can “see” at once.

For GPT-4-turbo, the window can handle up to 128,000 tokens, which sounds huge – but even that gets filled fast. You can’t keep the whole conversation there forever.

Instead, you use a vector store as long-term memory. You embed and save useful info.

Then, when needed, you query the vector store, retrieve the top relevant pieces, and feed them back into the LLM. This way, the model remembers just enough to act smart – without holding everything in its short-term memory.

Popular Vector Stores

There are several popular vector databases in use. Each one has its strengths.

FAISS (Facebook AI Similarity Search)

FAISS is an open-source library developed by Meta. It’s fast and works well for local or on-premise applications.

FAISS is great if you want full control and don’t need cloud hosting. It supports millions of vectors and provides tools for indexing and searching with high performance.

Here’s how you can use FAISS:

from sentence_transformers import SentenceTransformer
import faiss
import numpy as np

# Load a pre-trained sentence transformer model that converts sentences to numerical vectors (embeddings)
model = SentenceTransformer('all-MiniLM-L6-v2')

# Define the input sentence we want to store in memory
sentence = "User lives in Austin, Texas"

# Convert the sentence into a dense vector (embedding)
embedding = model.encode(sentence)

# Get the dimensionality of the embedding vector (needed to create the FAISS index)
dimension = embedding.shape[0]

# Create a FAISS index for L2 (Euclidean) similarity search using the embedding dimension
index = faiss.IndexFlatL2(dimension)

# Add the sentence embedding to the FAISS index (this is our "memory")
index.add(np.array([embedding]))

# Encode a new query sentence that we want to match against the stored memory
query = model.encode("Where is the user from?")

# Search the FAISS index for the top-1 most similar vector to the query
D, I = index.search(np.array([query]), k=1)

# Print the index of the most relevant memory (in this case, only one item in the index)
print("Most relevant memory index:", I[0][0])

This code uses a pre-trained model to turn a sentence like “User lives in Austin, Texas” into an embedding.

It stores this embedding in a FAISS index. When you ask a question like “Where is the user from?”, the code converts that question into another embedding and searches the index to find the stored sentence that’s most similar in meaning.

Finally, it prints the position (index) of the most relevant sentence in the memory.

FAISS is efficient, but it’s not hosted. That means you need to manage your own infrastructure.

Pinecone

Pinecone is a cloud-native vector database. It’s managed for you, which makes it great for production systems.

You don’t need to worry about scaling or maintaining servers. Pinecone handles billions of vectors and offers filtering, metadata support, and fast queries. It integrates well with tools like LangChain and OpenAI.

Here’s how a basic Pinecone setup works:

import pinecone
from sentence_transformers import SentenceTransformer

# Initialize Pinecone with your API key and environment
pinecone.init(api_key="your-api-key", environment="us-west1-gcp")

# Connect to or create a Pinecone index named "memory-store"
index = pinecone.Index("memory-store")

# Load a pre-trained sentence transformer model to convert text into embeddings
model = SentenceTransformer('all-MiniLM-L6-v2')

# Convert a fact/sentence into a numerical embedding (vector)
embedding = model.encode("User prefers vegetarian food")

# Store (upsert) the embedding into Pinecone with a unique ID
index.upsert([("user-pref-001", embedding.tolist())])

# Encode the query sentence into an embedding
query = model.encode("What kind of food does the user like?")

# Search Pinecone to find the most relevant stored embedding for the query
results = index.query(queries=[query.tolist()], top_k=1)

# Print the ID of the top matching memory
print("Top match ID:", results['matches'][0]['id'])

Pinecone is ideal if you want scalability and ease of use without managing hardware.

Other popular vector stores include:

• Weaviate – Combines vector search with knowledge graphs. Offers strong semantic search with hybrid keyword support.

• Chroma – Simple to use and good for prototyping. Often used in personal apps or demos.

• Qdrant – Open-source and built for high-performance vector search with filtering.

Each of these has its place depending on whether you need speed, scale, simplicity, or special features.

Making AI Seem Smart with Retrieval-Augmented Generation

This whole system – embedding user inputs, storing them in a vector database, and retrieving them later – is called retrieval-augmented generation (RAG).

The AI still doesn’t have a brain, but it can act like it does. You choose what to remember, when to recall it, and how to feed it back into the conversation.

If the AI helps a user track project updates, you can store each project detail as a vector. When the user later asks, “What’s the status of the design phase?” you search your memory database, pull the most relevant notes, and let the LLM stitch them into a helpful answer.

The Limits of Vector-Based Memory

While vector stores give AI agents a powerful way to simulate memory, this approach comes with some important limitations.

Vector search is based on similarity, not true understanding. That means the most similar stored embedding may not always be the most relevant or helpful in context. For instance, two sentences might be mathematically close in vector space but carry very different meanings. As a result, the AI can sometimes surface confusing or off-topic results, especially when nuance or emotional tone is involved.

Another challenge is that embeddings are static snapshots. Once stored, they don’t evolve or adapt unless explicitly updated. If a user changes their mind or provides new information, the system won’t "learn" unless the original vector is removed or replaced. Unlike human memory, which adapts and refines itself over time, vector-based memory is frozen unless developers actively manage it.

There are a few ways you can mitigate these challenges.

One is to include more context in the retrieval process, such as filtering results by metadata like timestamps, topics, or user intent. This helps narrow down results to what’s truly relevant at the moment.

Another approach is to reprocess or re-embed older memories periodically, ensuring that the information reflects the most current understanding of the user’s needs or preferences.

Beyond technical limitations, vector stores also raise privacy and ethical concerns. Key questions are: Who decides what gets saved? How long should that memory persist? And does the user have control over what is remembered or forgotten?

Ideally, these decisions should not be made solely by the developer or system. A more thoughtful approach is to make memory explicit. Let users choose what gets remembered. For example, by marking certain inputs as “important”, it adds a layer of consent and transparency. Similarly, memory retention should be time-bound where appropriate, with expiration policies based on how long the information remains useful.

Equally important is the ability for users to view, manage, or delete their stored data. Whether through a simple interface or a programmatic API, memory management tools are essential for trust. As the use of vector stores expands, so does the expectation that AI systems will respect user agency and privacy.

The broader AI community is still shaping best practices around these issues. But one thing is clear: simulated memory should be designed not just for accuracy and performance, but for accountability. By combining strong defaults with user control, developers can ensure vector-based memory systems are both smart and responsible.

Conclusion

Vector stores give AI agents a way to fake memory – and they do it well. By embedding text into vectors and using tools like FAISS or Pinecone, we give models the power to recall what matters. It’s not real memory. But it makes AI systems feel more personal, more helpful, and more human.

As these tools grow more advanced, so does the illusion. But behind every smart AI is a simple system of vectors and similarity. If you can master that, you can build assistants that remember, learn, and improve with time.

Hope you enjoyed this article. Connect with me on Linkedin.

Maybe you’ve said it yourself.

In the rush to launch a feature, meet a deadline, or impress a client, you take a shortcut. The code works – for now. The design passes – for now.

But over time, these choices pile up. They slow you down. They make every change harder. That’s technical debt.

Technical debt is a quiet, creeping cost. It doesn’t show up in metrics like churn or conversion rate. But it eats away at your product’s quality, your team’s velocity, and your ability to innovate.

You don’t notice it right away. Then, suddenly, everything feels slower. Nothing is simple anymore.

One bug fix breaks two new things. Engineers groan when they touch certain parts of the code. That’s when you realize you’re in debt.

Let’s talk about what technical debt really is, how it forms, and how you can deal with it before it kills your product.

What Is Technical Debt?

Think of technical debt like financial debt. When you borrow money, you get something now – at the cost of interest later.

In software, it’s the same. You make a quick decision to save time now, knowing it might cost you more effort down the line.

There’s nothing inherently wrong with that. Sometimes, taking on debt is smart. You might need to ship fast to test an idea or respond to the market.

But if you keep borrowing and never repay, the interest builds. And technical debt doesn’t just grow – it compounds. The longer you leave it, the worse it gets.

You don’t just end up paying more later. You slow down your whole team.

Every new feature takes longer. Bugs multiply. Morale drops. And eventually, your product feels fragile. That’s when the real damage begins.

Types of Technical Debt

Not all debt is the same. Some is like a short-term loan – you know you’re taking it and why. Other debt is like a bad mortgage – no one even knows it’s there until things break.

Here are the most common types:

• Intentional debt — You cut corners to hit a deadline but log it and plan to fix it. This can be healthy if managed well.

• Unintentional debt — You didn’t realise the shortcut was harmful. Often happens with new tech or unclear requirements.

• Environmental debt — Your tools, libraries, or frameworks get outdated. Even if your code is clean, it sits on crumbling infrastructure.

• Process debt — The way you build software becomes inefficient. Poor handoffs, unclear documentation, or weak testing pipelines all contribute.

Recognising which type you’re dealing with helps you prioritise. Not all debt needs immediate repayment. But all of it needs attention.

How Technical Debt Happens

As you can probably imagine, technical debt shows up in many ways.

Sometimes it’s deliberate. You make a tradeoff. You know it’s a shortcut, and you plan to clean it up later. That’s manageable – if you actually clean it up.

But most technical debt isn’t planned. It sneaks in through decisions that feel small in the moment. A rushed feature. A new hire not trained on the codebase. A spec that changes mid-sprint. Over time, these small cracks become deep fractures.

Here’s a simple example:

// Temporary workaround for product discounts
function applyDiscount(price, productType) {
  if (productType === 'electronics') {
    return price * 0.9; // 10% off
  } else if (productType === 'clothing') {
    return price * 0.8; // 20% off
  } else if (productType === 'books') {
    return price * 0.95; // 5% off
  } else {
    return price;
  }
}

This started as a quick fix. But over time, new product types get added with more exceptions.

Soon, you will have twenty if-else branches. It’s fragile. Every change risks breaking something. That’s technical debt.

The worst part? You might not even notice until a year later, when a bug in that logic takes hours to trace. You wonder, “How did this get so messy?” The answer: one shortcut at a time.

A better long-term approach in the above example would be a configuration-driven system or a discount rules engine.

// Config-driven discount logic
const discountRates = {
  electronics: 0.10,
  clothing: 0.20,
  books: 0.05
};

function applyDiscount(price, productType) {
  const discount = discountRates[productType] || 0;
  return price * (1 - discount);
}

Why Technical Debt Can Be Dangerous

Technical debt slows you down. That’s its most visible cost.

A feature that should take a day now takes a week. Simple changes break unrelated things. Your team spends more time fixing than building.

But the real danger goes deeper. Technical debt makes you afraid to touch your code.

Engineers stop refactoring because “it’s too risky.” You start saying no to new ideas because the system can’t handle them. The product becomes rigid. You stop innovating.

It also hurts your team. Developers don’t like working in messy codebases. It leads to burnout. New hires struggle to onboard.

Your best engineers spend their time firefighting instead of creating. Eventually, people leave. And your debt remains.

How to Manage Technical Debt

You can’t eliminate all technical debt. But you can manage it.

First, treat it like real debt. Track it. Prioritize it. Make regular payments.

Start by writing it down. Every time someone takes a shortcut, log it. You don’t need a fancy tool – a shared doc or Jira tag works fine. Just make it visible.

Next, build time into your workflow to pay it off. Use 10–20% of each sprint to refactor or improve the codebase. Don’t wait for a rewrite. Small, steady work adds up.

Code reviews help too. Encourage your team to ask: “Is this a shortcut?” If yes, make a conscious choice. Leave a clear comment. Note the tradeoff. Now it’s not a hidden cost – it’s a known one.

And when you do pay off debt, celebrate it. Make it part of your culture. The same way you’d celebrate shipping a feature, acknowledge when your team makes the codebase better. That builds pride and ownership.

Knowing When to Refactor

You can’t fix all the debt at once. So how do you choose?

Look for signs of pain. If a file breaks every sprint, fix it. If one part of the system takes days to test, improve it. If new hires always get stuck in one module, clean it up.

Focus on the code you touch often. There’s no point polishing a dead feature. But if something is part of your core flow, invest in it.

Also, listen to your team. Engineers know where the pain lives. If someone says, “This part scares me,” take that seriously. Fear in the codebase is a red flag.

When Debt Becomes Fatal

Sometimes, the debt gets so bad that small fixes won’t save you. The system collapses under its own weight. Everything feels slow. Nothing is safe to change. That’s when teams start talking about rewrites.

But rewrites are risky. They take time. They often miss hidden business logic. And they can carry old debt into new code if not done carefully.

If you must rewrite, do it incrementally. Replace modules one at a time. Add tests. Migrate data with care.

And don’t forget why the old system failed. If you don’t fix the culture, the new system will rot too.

Final Thoughts

Technical debt isn’t evil. It’s part of building software. But like financial debt, it needs discipline. You can’t just ignore it and hope it goes away.

Great products aren’t just well-designed. They’re well-maintained. The teams behind them care about quality — not just in what users see, but in what engineers live with every day.

So the next time you say, “We’ll fix it later,” ask yourself: will you? Or are you just borrowing against the future?

Hope you enjoyed this article. Join my newsletter for similar articles and connect with me on Linkedin.

It’s not just about learning a programming language and typing out code to build software. There’s a lot more to it. And one of the most confusing (and often frustrating) topics for developers is clean code.

So, what is clean code?

Simply put, it’s about writing code that’s so clear and well-organized that neither you nor anyone else will get frustrated trying to understand it six months later.

Think of clean code as the programming equivalent of great design — it’s functional, beautiful, and easy to work with.

And today, I won’t spend a lot of time explaining why clean code is important — you probably already know that. Instead, I’ll get straight to the point and share seven powerful hacks to help you write cleaner, better code.

Table of Contents

• 1. Write Code Like You’re Explaining it to a 5-Year Old

• 2. Use AI Tools (or an AI Code Reviewer)

• 3. Get Rid of Unnecessary Comments

• 4. Follow the DRY Principle

• 5. Fix Your Code Formatting & Follow a Consistent Style

• 6. Don’t Let Your Functions Do Too Much

• 7. Organize Your Files and Folders Properly

• Wrapping Up

1. Write Code Like You’re Explaining it to a 5-Year Old

Let me be honest — if you are writing exceedingly clever code that your teammates or someone else can’t easily read, it won’t be helpful to anyone.

You need to write code so simple that anyone, including someone who just opened the file for the first time, can easily go through it.

For example, if your variable names look like this:

let x = y + z;

This isn’t helpful. No one will know what x, y, and z mean — not even you, three weeks later.

Variables should describe what they hold. Think of them as self-documenting comments. Here’s a better example:

let totalPrice = productPrice + shippingCost;

This simple best practice can be applied even when writing functions, comments, and more.

Here’s an example of hard to understand code:

function calc(itm) {
 let t = 0;
 for (let i = 0; i < itm.length; i++) {
 t += itm[i].p;
 }
 return t;
}

You see, it doesn’t give you a proper idea about the function other than the logic.

Instead, try to write it like this:

function calculateTotalPrice(cartItems) {
    let totalPrice = 0;
    for (let i = 0; i < cartItems.length; i++) {
        totalPrice += cartItems[i].price;
    }
    return totalPrice;
}

Now it’s clear just from looking at the code what’s going on.

2. Use AI Tools (or an AI Code Reviewer)

AI is evolving quickly and is being used in almost every industry.

Well, as a developer, you can use AI to help you write clean and readable code — all thanks to AI tools like ChatGPT, Claude, and GitHub Copilot.

For starters, you can copy and paste your code into an LLM and ask it to review your code.

Here’s an example of a request I made to ChatGPT:

And here’s what ChatGPT recommended:

It even provided me with an improved version of the code. Here it is as follows:

Along with that, you can also use AI-powered code reviewers like CodeRabbit AI, which integrates with your pull requests to offer automated code reviews, walkthroughs, and more.

In simple terms, when you install CodeRabbit AI, add it to your pull request, and then create a pull request, CodeRabbit AI provides you with summaries, code reviews, walkthroughs, and more.

Here’s an example from an open-source repository “minefield” of a summary generated by CodeRabbit AI:

In the above examples, CodeRabbit AI simply goes through the code and provides a great summary highlighting the new features, bug fixes, and unit tests added via a pull request.

If you want to learn more, here’s a quickstart guide that provides more info on how to get started and use CodeRabbit AI.

Note that, AI tools can be a great help in improving your code, but they should never replace your own thinking.

When AI suggests changes, always ask yourself: Does this actually make sense?

After all, AI isn’t perfect—it can generate incorrect code. It might also miss important details that only a human can fully understand.

So instead of blindly accepting AI’s suggestions, take a moment to understand why they were made.

Remember, the goal is to use AI to speed up your workflow, catch errors, and more—while keeping full control over your code.

3. Get Rid of Unnecessary Comments

Writing good code comments can help others understand why your code is doing what it’s doing.

But I see developers writing too many comments, even where there is no need.

I believe that, whenever possible, good code should document itself.

Here’s an example of a not so helpful comment:

// Adding 10 to the result 
total = total + 10;

You see, here the comment doesn’t make sense, and it doesn’t really add anything to the code (or our understanding of it).

It’s a better practice to write comments only if they’re important in helping a reviewer understand why you did something in particular or if there’s some ambiguity that needs explaining.

Here’s an example:

// Adding 10 because the client requires a 10% buffer for calculations 
total = total + 10;

Now you can see that here the comment gives a clear idea about why the programmer added 10 to the total.

4. Follow the DRY Principle

I’ve seen a lot of programmers repeating the same logic or adding the same functionality in different files.

Well, you don’t need to repeat the same code everywhere because it makes the process much more complex. Keep your code DRY (Don’t Repeat Yourself).

Instead, abstract that logic into a single reusable function.

For example, instead of writing the same logic in different files:

if (user.age > 18 && user.age < 65) { 
 // Do something 
}

if (user.age > 18 && user.age < 65) { 
 // Do something else 
}

You can create a reusable function so that you can use the logic everywhere, as shown below:

function isWorkingAge(age) { 
 return age > 18 && age < 65; 
}

if (isWorkingAge(user.age)) { 
 // Do something 
}

In short — write once, and use everywhere.

5. Fix Your Code Formatting & Follow a Consistent Style

This is another simple hack, but many devs don’t even think about it.

First of all, you need to format your code with proper indentation.

You can just install a VS Code extension/linters like Prettier, ESLint, or Flake8, depending on the programming language you use. Configure a few settings, and you’re good to go.

These linters can help you write better code by finding mistakes, helping you follow coding rules, and keeping your code consistent. They can also catch errors, make your code easier to read, and save time on fixing bugs and reviews.

But fixing your formatting doesn’t just mean it’s clean code – it’s way more than that.

Beyond formatting, you should stick to a consistent style guide for things like function names or variable names.

For example, here’s some code that has an inconsistent style:

let total_price;  
let UserData;  
function getuser() {}

But some of you may ask, why is the code inconsistent?

Well, it uses several different naming conventions – like total_price uses snake_case, UserData uses PascalCase, and getuser() is in lowercase instead of camelCase.

This makes your code harder to read and more confusing. Instead, you can follow a consistent style. Here’s how:

let userData; 
let totalPrice; 
function getUser() {}

And just to let you know, JavaScript typically follows camelCase for variables and functions, so names like getUser(), userData, and totalPrice would be more consistent.

No matter whether you’re working in a team or solo, sticking to one style is a good idea. It makes the code clean, and the reviewer can easily go through it.

6. Don’t Let Your Functions Do Too Much

As a programmer, I know that sometimes you need to write complex logic inside a function.

But most often, programmers include too much logic in a single function, causing it to do more than one thing at a time. These functions become too complex, making them hard to read or understand.

It’s better to create multiple smaller functions, with each function handling a single responsibility.

Here’s an example of a slightly too complex function:

function calculateCart(items) {
 let total = 0;
 for (let i = 0; i < items.length; i++) {
 total += items[i].price * items[i].quantity;
 }
 return total > 100 ? total * 0.9 : total;
}

You see, this single function does too much—it applies discounts if applicable, and calculates the total price based on quantity, making it hard to reuse for specific use cases.

A better way is to split it into three functions:

• One to apply the discount (if applicable).

• One to calculate the total price based on quantity.

• One to get the total and apply the discount if needed.

This makes the code cleaner and easier to manage. Here’s an improved version:

function calculateCart(items) {
 const total = getCartTotal(items);
 return applyDiscount(total);
}

function getCartTotal(items) {
 return items.reduce((sum, item) => sum + item.price * item.quantity, 0);
}

function applyDiscount(total) {
 return total > 100 ? total * 0.9 : total;
}

You see, instead of writing everything in one long block, the logic is broken into smaller, clearer functions.

And that’s what makes it easier for you (and everyone else) to understand.

7. Organize Your Files and Folders Properly

Now comes the most important part.

When you’re working on a big project, it makes no sense to dump all your code into one single folder with a bunch of random files. It’s become a nightmare for anyone trying to review the code or find a specific file.

Think about it — going through 50 files just to fix a bug. Nobody wants that.

Instead, try organizing your project into multiple folders based on pages or features.

And don’t stop there — break large files into smaller, specific modules based on functionality. This way, everything is neat, easy to find, and makes sense at first glance.

Here’s a bad example of a project structure:

project-folder/  
│  
├── index.html  
├── app.js  
├── helpers.js  
├── data.js  
├── user.js  
├── product.js

It’s messy, right? Now, here’s how you can improve it:

project-folder/  
│  
├── pages/  
│   ├── home/  
│   │   ├── HomePage.js  
│   │   ├── HomePage.css  
│   │   └── HomePage.test.js  
│   ├── user/  
│   │   ├── UserPage.js  
│   │   ├── UserPage.css  
│   │   └── UserPage.test.js  
│   └── product/  
│       ├── ProductPage.js  
│       ├── ProductPage.css  
│       └── ProductPage.test.js  
├── components/  
│   ├── Header.js  
│   ├── Footer.js  
│   └── Button.js  
├── utils/  
│   └── api.js  
└── index.js

See how simple this is?

You have separate folders for pages, components, and utilities. Inside each folder, files are organized by purpose and given clear names.

For example, if you need something for the product page, you’ll know exactly where to find it.

Wrapping Up

In this article, you learned the basics you need to get started writing clean, efficient, and maintainable code.

We discussed how to write clean code, use AI tools, apply proper code formatting techniques, structure functions effectively, and follow other best practices.

If you apply these tips consistently, you’ll significantly improve your coding skills and write cleaner code.

I hope you liked it.

You can connect with me on Substack, and Twitter.

Also, if you're interested in learning more about AI, you can subscribe to my newsletter, AI Made Simple, where I dive deeper into practical AI strategies for everyday people.

What is a Framework?

In terms of software development, a framework is a structured set of tools, libraries, and conventions that provides a foundation for building applications more efficiently by handling common functionalities so developers can focus on unique features.

A modern web application can combine multiple frameworks to handle different aspects of development efficiently:

• React with Tailwind CSS manages the user interface and styling.

• FastAPI or Django with LangChain handle backend operations and AI functionality, while MongoDB acts as a memory store.

These frameworks communicate through APIs and defined interfaces to work together—for example, a user interaction in the React front end can trigger an AI process through LangChain in the Python back end, which uses MongoDB Atlas Vector Search to retrieve relevant data and then display it back to the user with Tailwind CSS styles.

Table of Contents:

• The framework landscape in 2024

• Application frameworks

• AI frameworks

• Web frameworks

• CSS/UI frameworks

• Testing & infrastructure

• Making the right choice

• The bottom line

• Additional resources

• Frequently asked questions (FAQ)

The Framework Landscape in 2024

Think of frameworks as tools in a technical toolbox. Understanding what each tool is designed to do—and what it isn't—ensures you use the right solution for the task at hand.

The Evolution of Development Frameworks

The framework landscape has transformed dramatically over the past decade. What started as simple libraries for rendering web pages has evolved into sophisticated ecosystems that can:

• Handle complex state management.

• Process real-time data streams.

• Integrate AI capabilities.

• Scale automatically based on demand.

• Deploy across multiple platforms from a single codebase.

Why Framework Choice Matters

Frameworks are puzzle pieces: each has a unique shape and function. When chosen wisely, they fit together seamlessly to create a cohesive application. But forcing incompatible pieces together can lead to inefficiencies and broken functionality.

Here's why your choice matters:

Technical impact

• Performance: Different frameworks have different performance characteristics. Instagram chose React for its virtual DOM, which handles frequent updates efficiently.

• Scalability: Uber's back end uses Node.js because it excels at handling numerous concurrent connections.

• Maintenance: Shopify standardized on React Native to maintain a single codebase for mobile applications.

Business impact

• Development speed: The right framework can accelerate development by 2-3x.

• Team productivity: Familiar frameworks reduce onboarding time from months to weeks.

• Cost efficiency: Proper framework selection can significantly reduce hosting and maintenance costs.

Common pitfalls

• Over-engineering: Using Next.js when a simple HTML page would suffice

• Under-engineering: Using vanilla JavaScript for a complex, state-heavy application

• Misaligned tools: Using Electron (a desktop app framework) to build a simple website

• Trend-chasing: Adopting the newest framework without considering maintenance implications

Let’s dive deeper into the various categories of development frameworks and how they can help you build better applications.

Note: I give framework examples throughout this article, but these are not exhaustive lists and new frameworks are created daily|weekly|monthly.

Application Frameworks

These are your Swiss Army knives of development—comprehensive toolkits that handle the entire application lifecycle. While each framework has its specialties, they typically provide:

• Database integration and ORM support.

• Authentication and authorization.

• API routing and middleware.

• Template engines or component systems.

• Asset management.

• Security features.

• Development tools and debugging support.

What is an application? An application is the combination of individual components working together, including a user interface and backend services, to perform specific functions or a set of functions for users. It’s designed to be fully deployable and operate in production environments.

You'll encounter them in three main flavors:

Full-Stack Web Application Frameworks

• Django and Rails: Perfect for data-heavy applications with complex business logic. They follow the "batteries included" philosophy, providing everything you need out of the box.

• Next.js and Nuxt.js: Modern full-stack frameworks optimized for React and Vue.js, respectively. They excel at building performant applications with capabilities like server-side rendering, static site generation, and API integration.

• Spring Boot: Enterprise-grade framework favored for large-scale Java applications, particularly in financial and banking sectors.

Mobile Application Frameworks

• Flutter: Google's toolkit for building natively compiled apps for mobile, web, and desktop from a single codebase. Known for smooth animations and native performance.

• React Native: Ideal when you want to leverage your team's React knowledge for mobile development. Great for apps that need to feel native while maintaining code sharing between platforms.

• SwiftUI and Jetpack Compose: Native frameworks for iOS and Android, respectively. Ideal when platform-specific features and optimal performance are crucial.

Desktop Application Frameworks

• Electron: Powers apps like VS Code and Slack. Ideal for leveraging existing web development knowledge without having to learn OS-specific languages to build cross-platform desktop applications.

• Tauri: A modern alternative to Electron with smaller bundle sizes and better performance. Uses native system webviews to run apps more efficiently, with less memory and faster startup times. Developers can use their existing web development skills without learning new programming languages for each OS.

• PyQt: A good option for building GUI applications in Python, especially for data science projects. It has a rich set of libraries and widgets, suitable for both small tools and complex apps.

Gaming Frameworks

• Unity: A popular framework used for creating 2D and 3D games. It supports multiple platforms, making it a versatile choice for game developers. Unity is known for its user-friendly interface and extensive asset store.

• Unreal Engine: Known for its high-quality graphics, Unreal Engine is a powerful framework used for both AAA games and indie projects. It provides advanced visual tools and real-time rendering capabilities.

• Godot: An open-source game engine that's lightweight and flexible. Godot is often used for smaller or indie game projects and features an intuitive scene system that makes development straightforward.

AI Frameworks

Here's a crucial point many miss: AI frameworks like TensorFlow or LangChain are powerful, but they're not standalone solutions. These frameworks require integration with other tools and frameworks for data handling, user interface, and deployment to create a complete, production-ready application. These are just another piece of the puzzle.

The exception? Data scientists can use these tools directly in Jupyter notebooks for research and prototyping. But for production applications, you'll need more.

Let's look at the major players and their sweet spots:

AI Model Development Frameworks

• TensorFlow: Google's powerhouse for deep learning, TensorFlow is perfect for applications involving computer vision, neural networks, and production-grade ML. Pinterest uses it for image recognition and recommendations, highlighting its strength in processing large amounts of image data efficiently.

• PyTorch: Developed by Facebook, PyTorch is a flexible framework ideal for research, natural language processing, and quick prototyping. Tesla has utilized PyTorch for various machine learning tasks, including research related to autonomous driving vision systems, demonstrating its versatility in cutting-edge applications.

• JAX: A high-performance numerical computing framework by Google, JAX is well-suited for scientific computing and large-scale transformers. DeepMind has used JAX for advanced AI research, including projects like AlphaFold for protein structure prediction, showcasing its effectiveness in large-scale computational problems.

AI Model Deployment and Serving Frameworks

• TensorFlow Serving: A deployment tool for TensorFlow models, perfect for serving high-performance ML models in production environments. For example, it has been used to serve image classification models for e-commerce platforms, ensuring rapid and scalable response times.

• TorchServe: A robust tool for deploying PyTorch models, perfect for scalable PyTorch model serving. It has been used for deploying chatbot models for real-time customer support, providing flexibility and efficiency in managing conversational AI.

• NVIDIA Triton Inference Server: A multi-framework model serving tool, NVIDIA Triton Inference Server is capable of handling models from TensorFlow, PyTorch, ONNX, and more. It is ideal for managing inference requests for multi-modal AI applications, making it highly suitable for complex AI deployments.

LLM Integration Frameworks

• LangChain: The Swiss Army knife for LLM applications. Ideal for building chatbots and document Q&A systems. It has been employed to create customer service bots capable of accessing company knowledge bases, demonstrating its usefulness in enhancing customer interactions.

• LlamaIndex: Specializing in data connection, LlamaIndex is perfect for building search engines over private data. It has been used to create semantic search systems for internal documents, facilitating more effective data retrieval and organizational knowledge management.

• Hugging Face Transformers: A pre-trained model hub, Hugging Face Transformers allows for the quick deployment of state-of-the-art models. It has been used to add sentiment analysis to customer feedback systems, highlighting its capability in natural language understanding.

AI Data Processing Frameworks

• Apache Spark (MLlib): Ideal for large-scale data transformation and ML, used for processing millions of user interactions in recommendation systems. Its scalability and efficiency make it a popular choice for big data ML tasks.

• Pandas: A widely-used data manipulation and analysis tool that is perfect for data cleaning, analysis, and feature engineering. It is often employed in preparing customer data for churn prediction models, owing to its intuitive data handling capabilities.

• Polars: A high-performance data manipulation framework, used when Pandas is not fast enough for processing requirements. It has been applied in real-time financial data analysis, delivering faster data processing and efficiency.

AI Automation and Orchestration Frameworks

• CrewAI: Used to orchestrate multiple AI agents, making it perfect for complex workflows requiring multiple AI models. It has been used to create a content creation pipeline that plans, writes, and edits, showcasing its ability to automate creative processes.

• Auto-GPT: Develops autonomous AI agents, ideal for self-directed task completion. It has been utilized for automated research and data gathering, highlighting its potential for automating repetitive knowledge tasks.

• Microsoft Semantic Kernel: An AI orchestration tool perfect for integrating AI into .NET applications. It has been used to add AI capabilities to existing enterprise applications, providing a seamless integration of AI functions into established workflows.

AI Data Ingestion and Document Processing Frameworks

• Apache NiFi: A powerful tool for data flow automation, well-suited for real-time data ingestion and processing. It has been used to extract and transform log data before analysis, ensuring efficient data flow for various use cases.

• Haystack: Specialized in document processing for search and QA systems. It has been utilized to create pipelines for indexing documents and answering questions, making it ideal for building internal knowledge base searches.

• Unstructured: Designed for extracting data from a variety of formats, including PDFs, HTML, and images. It’s perfect for handling unstructured content in document processing workflows and has been used to extract relevant information from scanned documents for data analysis.

• Airbyte: An open-source data integration tool that is ideal for connecting and syncing data across multiple sources. It has been used to ingest data from third-party APIs into analytics systems, enabling effective data consolidation.

Interactive AI Web Frameworks

• Gradio: Simplifies the process of building web UIs for machine learning models, making it perfect for creating quick demos for ML models. It has been used to create interactive image classifiers for end users, providing an accessible interface for testing ML capabilities.

• Streamlit: A Python-based web app framework for ML and data science, ideal for turning data scripts into shareable web apps. It has been used to build user-friendly dashboards for exploring model predictions, enhancing the accessibility of ML models.

MLOps and AI Monitoring Frameworks

• MLflow: Used for experiment tracking and model lifecycle management, perfect for keeping track of experiments and model versions. It has been applied to manage multiple iterations of predictive models, supporting organized development workflows.

• Kubeflow: A Kubernetes-native platform for MLOps, ideal for deploying, scaling, and managing machine learning models on Kubernetes. It has been used to run end-to-end ML workflows in production environments, ensuring scalability and consistency.

• Prometheus & Grafana: Monitoring and alerting tools for ML infrastructure, perfect for tracking model inference performance and system metrics. They have been used to monitor latency and resource usage of deployed ML services, ensuring optimal operational performance.

Web frameworks

Web development has evolved into two distinct camps—think of them as the front-of-house and kitchen staff in a restaurant. Both essential, both specialized, but with very different responsibilities.

Note: We already discussed full-stack web application frameworks above, which marry these two together.

Frontend Frameworks

Frontend frameworks handle what users see and interact with, managing everything from data display to user interactions, user input, and how the overall user experience is structured. These frameworks ensure that the interface is visually appealing, intuitive, and responsive to user actions.

They play a critical role in how smoothly data is presented, processed, and updated in real time, providing dynamic elements like animations, form validation, and client-side routing to enhance usability.

By using frontend frameworks, developers can create highly interactive and cohesive user experiences that feel natural and engaging.

• React: Ideal for large-scale applications with complex state management. It features a virtual DOM for optimal performance, making it suitable for highly interactive user interfaces.

• Vue.js: Perfect for both small projects and enterprise applications. It has a gentle learning curve combined with powerful scalability, making it an approachable yet robust frontend solution.

• Svelte: Ideal for when you have performance-critical applications and want smaller bundles. It compiles away framework code for lighter applications, providing better performance and a smaller footprint.

Backend Frameworks

Backend frameworks manage server-side logic, data processing, and system integration, handling everything from receiving and processing client requests to interacting with databases and external APIs. These frameworks ensure that the server processes are efficient, scalable, and secure, supporting high concurrency and maintaining consistent data flows.

They provide the essential tools for developers to build, maintain, and optimize the server-side of applications, including tasks such as handling authentication, managing business logic, and ensuring data consistency.

A well-chosen backend framework allows developers to focus more on creating features rather than dealing with low-level server management.

• Express.js: Perfect for APIs and microservices, offering minimal structure with maximum flexibility. It is highly popular in Node.js environments for building efficient server-side logic.

• FastAPI: Designed for high-performance APIs, with automatic API documentation and type checking. It is commonly used for fast and secure backend implementations in Python.

• NestJS: For large-scale Node.js applications, featuring an Angular-inspired architecture that enhances scalability. It provides a well-structured framework for enterprise-level backends.

CSS/UI Frameworks

Modern CSS frameworks have evolved beyond simple styling to become complete design systems. Gone are the days when CSS frameworks just provided basic grid systems and button styles.

Today's frameworks are sophisticated tools that enable consistent design at scale. They offer features such as responsive layouts, dark mode support, accessibility enhancements, and interactive components.

They've become essential for maintaining design consistency across large applications and teams, while significantly reducing development time and technical debt. Some frameworks even include built-in performance optimization, design tokens for brand customization, and tools for managing design systems across multiple platforms.

They fall into two main categories:

Utility-first Frameworks

• Tailwind CSS: Custom designs with consistent systems. It features rapid prototyping capabilities, highly customizable design tokens, and small production bundles through PurgeCSS.

• UnoCSS: A customizable, efficient, utility-first CSS solution. It provides on-demand CSS for optimal performance and minimal bundle size, making it a modern tool for design system management.

Component-based Frameworks

• Bootstrap: Rapid prototyping and traditional web applications. It offers an extensive library of pre-built components, a responsive grid system, and a rich ecosystem of themes and plugins.

• Material UI: For applications following Material Design guidelines. It includes a comprehensive component library, a theming system, and accessibility compliance features.

Testing and Infrastructure

These frameworks form the backbone of reliable, scalable applications. While they might not get as much attention as flashy frontend frameworks or AI tools, they're the critical infrastructure that keeps modern applications running smoothly at scale.

Testing frameworks ensure code quality and prevent regressions, while containerization and orchestration frameworks handle the increasingly complex world of cloud deployment and scaling.

In an era where a single code change can affect millions of users instantly, or where an application might need to scale from hundreds to millions of requests in minutes, these frameworks aren't just nice-to-have—they're essential for survival. Major tech companies like Netflix, Amazon, and Google rely on these tools to maintain their rapid development pace while ensuring reliability and performance at massive scale.

Testing Frameworks

• Jest: Features snapshot testing, code coverage, and mocking capabilities, making it a popular choice for ensuring the quality of JavaScript codebases.

• Cypress: Provides real browser testing and time-travel debugging, ideal for end-to-end testing of web applications.

• PyTest: Known for its simple syntax, powerful extensions, and easy fixture management, making it a go-to framework for testing Python applications.

Containerization and Orchestration

• Docker: Used for application containerization, providing consistent development and deployment environments with isolated dependencies and efficient resource usage.

• Kubernetes: Orchestration of large-scale distributed applications. It provides automated scaling, self-healing deployments, and rolling updates, ensuring that containerized applications can grow as needed.

Making the Right Choice

Selecting the right framework can determine the success of your entire project. Consider these critical points to guide your decision-making process:

1. Application Type and Scale

• Small site? Consider React or Vue.js.

• Large application? Next.js or Django might be better.

• Need SEO? Look for SSR capabilities.

2. Team Capabilities

• Strong JavaScript? Consider Node.js ecosystem.

• Python experts? Django or FastAPI might be better.

• Need quick ramp-up? Combine key frameworks like Vue.js and Bootstrap.

3. Technical Requirements

• High-performance computing? Consider Rust frameworks.

• Real-time updates? Look at WebSocket support.

• AI requirements? Integration with platforms like Hugging Face and frameworks like LangChain might be essential.

4. Scaling Strategy

• Vertical scaling? Simpler frameworks might suffice.

• Horizontal scaling? You need frameworks with microservices support.

• Global distribution? Consider edge computing capabilities.

5. Long-term Maintenance

• Community size and activity

• Available talent pool

• Corporate backing and stability

• Documentation quality

• Upgrade path complexity

The Bottom Line

Framework selection isn't just a technical decision—it's a strategic one that affects your project's success, team productivity, and maintenance costs. So take the time to understand your options.

Now you should have all of the puzzle pieces, but it's up to you to put them together thoughtfully to create the perfect technology stack that meets both your current and future needs.

Additional Resources

If you'd like to learn how MongoDB integrates seamlessly with many of the frameworks mentioned in this guide, check out the resources below. You'll find that it offers a flexible and scalable database solution that adapts to different use cases.

• How to Integrate MongoDB Into Your Next.js App

• Introducing FARM Stack - FastAPI, React, and MongoDB

• How to Use MERN Stack: A Complete Guide

• Build a JavaScript AI Agent With LangGraph.js and MongoDB

• RAG with Atlas Vector Search, LangChain, and OpenAI

Frequently Asked Questions (FAQ)

What is a development framework?

A development framework is a set of tools, libraries, and conventions that provide a foundation for building applications efficiently. Frameworks handle common tasks so developers can focus on building unique features.

How do AI and development frameworks work together?

AI frameworks can be combined with traditional development frameworks to create intelligent, data-driven applications. For instance, a backend framework like FastAPI can handle incoming requests, while an AI framework like LangChain processes natural language data, providing users with enhanced functionality like chatbots or recommendation systems.

Why is choosing the right framework important?

Choosing the right framework can impact your project's performance, scalability, and maintenance. It ensures that your application meets user expectations, is easy to manage, and can grow as needed.

How do I decide which framework to use for my project?

Consider factors like application type, team expertise, performance needs, scalability, and long-term maintenance. Each of these can guide you in selecting a framework that matches your specific requirements.

Can I use multiple frameworks in one project?

Yes, combining multiple frameworks can be very effective. You might use one framework for the front end, another for backend services, and a third for AI integrations, ensuring that each part of your application is handled by the best-suited tool.

How does MongoDB fit into development frameworks?

MongoDB integrates with many frameworks, acting as the data layer for your applications. It provides a flexible, scalable solution that supports full-stack, backend, and AI frameworks to store and retrieve data efficiently.

The Discord API provides you with an easy tool to create and use your own bots and tools.

In this tutorial, you'll learn how you can create a basic music bot and add it to your server. The bot will be able to play, skip, and stop the music, and will also support queuing functionality.

Table of Contents

1. Prerequisites
2. How to Set Up a Discord Bot
   – How to add the bot to your server
   – How to create your project
   – Discord.js basics
3. Discord Bot Version 0.13
   – How to create the Discord player
   – How to add slash commands
   – How to implement interactions
   – How to play songs
   – How to skip songs
   – How to stop songs
   – Complete source code for index.js
4. Discord Bot Version 0.12
   – How to read messages
   – How to add songs
   – How to play songs
   – How to skip songs
   – How to stop songs
   – Complete source code for index.js
5. Conclusion

Prerequisites

Before we get started creating the bot, make sure that you have installed all the tools you'll need:

• Node
• NPM
• FFMPEG

After you've installed these, you can continue by setting up your discord bot.

How to Set Up a Discord Bot

First, you need to create a new application on the discord development portal.

You can do so by visiting the portal and clicking on New Application.

Create a new Discord application

After that, you need to give your application a name and click the Create button.

Give your bot whatever name you like - I've chosen "music-bot"

After that, select the bot tab and click on Add Bot.

Add your bot under the "Bot" tab

Now your bot is created and you can continue with inviting it to your server.

How to add the bot to your server

After creating your bot, you can invite it using the OAuth2 URL Generator.

For that, you need to navigate to the OAuth2 page and select bot in the scope tap.

Selecting "bot" on the 0Auth2 Generator page

After that, you need to select the needed permissions to play music and read messages.

Select the permissions you'll need - "read messages/view channels", "send messages", "manage messages", "add reactions", "use slash commands", "connect", and "speak.

Then you can copy your generated URL and paste it into your browser.

Copy the URL

After pasting it, add it to your server by selecting the server and clicking the authorize button.

How to create your project

Now you can start creating your project using the terminal.

First, create a directory and move into it. You can do so by using these two commands:

mkdir musicbot && cd musicbot

After that, create your project modules using the npm init command. After entering the command, you will be asked some questions – just answer them and continue.

Then you just need to create the two files you will work in.

touch index.js && touch config.json

Now, open your project in your text editor. I personally use VS Code and can open it with the following command:

code .

Discord.js basics

Now you need to install some dependencies before we can get started.

npm install discord.js@^12.5.3 ffmpeg fluent-ffmpeg @discordjs/opus ytdl-core --save

After the installation finishes, you can continue with writing your config.json file. Here, save the token of your bot and the prefix it should listen for.

{
"prefix": "!",
"token": "your-token"
}

To get your token, you need to visit the discord developer portal again and copy it from the bot section.

Get your bot token by clicking "Copy" and save it somewhere safe

Those are the only things you need to do in your config.json file. So now it's time to start writing your JavaScript code.

The article includes two versions: one for the new discord.js v13, which uses slash commands combined with the discord-player library to implement the music functionality, and one for discord.js v12.5.3, which implements the functionality without a library.

The older version is better for learning purposes, and the newer version works with the current discord.js and is a lot easier to implement – so choose which you prefer.

Discord Bot Version 0.13

Now you just need to install some more dependencies before we can get started.

npm install discord.js discord-player @discordjs/opus

After installing the dependencies, import them in your dependencies.

const { Client, GuildMember, Intents } = require("discord.js");
const { Player, QueryType } = require("discord-player");
const config = require("./config.json");

After that, create your client and log in using your token.

const client = new Client({
    intents: [Intents.FLAGS.GUILD_VOICE_STATES, Intents.FLAGS.GUILD_MESSAGES, Intents.FLAGS.GUILDS]
});
client.login(config.token);

Now add some basic listeners that console.log when they get executed.

client.once('ready', () => {
 console.log('Ready!');
});

client.on("error", console.error);
client.on("warn", console.warn);

After that, you can start your bot using the node command and the bot should be online on Discord and print “Ready!” in the console.

node index.js

How to create the Discord player

Now that you've created the client for the discord bot, you can continue by initializing your player. This will allow you to play and manage music in your Discord channel.

const player = new Player(client);

You can also add some error handlers that will be called if an error occurs.

player.on("error", (queue, error) => {
    console.log(`[${queue.guild.name}] Error emitted from the queue: ${error.message}`);
});
player.on("connectionError", (queue, error) => {
    console.log(`[${queue.guild.name}] Error emitted from the connection: ${error.message}`);
});

The last thing you need to do is add listeners for the different player events like a song starting or being added.

player.on("trackStart", (queue, track) => {
    queue.metadata.send(`🎶 | Started playing: **${track.title}** in **${queue.connection.channel.name}**!`);
});

player.on("trackAdd", (queue, track) => {
    queue.metadata.send(`🎶 | Track **${track.title}** queued!`);
});

player.on("botDisconnect", (queue) => {
    queue.metadata.send("❌ | I was manually disconnected from the voice channel, clearing queue!");
});

player.on("channelEmpty", (queue) => {
    queue.metadata.send("❌ | Nobody is in the voice channel, leaving...");
});

player.on("queueEnd", (queue) => {
    queue.metadata.send("✅ | Queue finished!");
});

In most cases, you just send a message into the Discord text channel using the send() function.

How to add slash commands

After you've set up the player successfully, you can continue by adding your Slash commands to your client. This step lets Discord know which commands the bot can execute.

client.on("messageCreate", async (message) => {
        if (message.author.bot || !message.guild) return;
    if (!client.application?.owner) await client.application?.fetch();
});

You can do this by implementing a simple !deploy command that saves your commands in the guild.commands variable of a message.

A slash command has a name, a description, and an optional options field that contains the command’s parameters. For example, the play command takes a song query as an argument.

client.on("messageCreate", async (message) => {
        ...

        if (message.content === "!deploy" && message.author.id === client.application?.owner?.id) {
        await message.guild.commands.set([
            {
                name: "play",
                description: "Plays a song from youtube",
                options: [
                    {
                        name: "query",
                        type: "STRING",
                        description: "The song you want to play",
                        required: true
                    }
                ]
            },
            {
                name: "skip",
                description: "Skip to the current song"
            },
            {
                name: "queue",
                description: "See the queue"
            },
            {
                name: "stop",
                description: "Stop the player"
            },
        ]);

        await message.reply("Deployed!");
    }
});

After entering !deploy in your Discord text chat, the slash commands will be added to your application. When typing / into the chat you should see something similar to this:

Example of using the slash commands

How to implement interactions

Once the interactions (slash commands) are defined, now you'll need to implement them.

All slash commands trigger the interactionCreate event and can be implemented inside the async function below. Before executing any functionality, run a few conditionals to check if the user is allowed to perform the given functionality.

client.on("interactionCreate", async (interaction) => {
    if (!interaction.isCommand() || !interaction.guildId) return;

    if (!(interaction.member instanceof GuildMember) || !interaction.member.voice.channel) {
        return void interaction.reply({ content: "You are not in a voice channel!", ephemeral: true });
    }

    if (interaction.guild.me.voice.channelId && interaction.member.voice.channelId !== interaction.guild.me.voice.channelId) {
        return void interaction.reply({ content: "You are not in my voice channel!", ephemeral: true });
    }
});

After that, check which command is being executed by matching the commandName with the name of the commands you defined above.

client.on("interactionCreate", async (interaction) => {
    ...

        if (interaction.commandName === "play") {
            // TODO: Implement play command
        }
});

You can then add the implementation inside of the if statement.

How to play songs

The play command requires you to search for the provided song and add the result to the current queue of songs.

Let’s start by retrieving the user-provided query using the options.get() function. After that you can use the player.search() function to search for the desired song.

if (interaction.commandName === "play") {
    await interaction.deferReply();

    const query = interaction.options.get("query").value;
    const searchResult = await player
        .search(query, {
            requestedBy: interaction.user,
            searchEngine: QueryType.AUTO
        })
        .catch(() => {});
    if (!searchResult || !searchResult.tracks.length) return void interaction.followUp({ content: "No results were found!" });
}

Now that you have the song, you can create a queue for the songs (if there is already a queue, the createQueue function will return the existing one).

Once the queue is created, you can try joining the user’s voice channel. If that is successful, add the song to the current queue using the addTracks function.

if (interaction.commandName === "play") {
    ...

        const queue = await player.createQueue(interaction.guild, {
        metadata: interaction.channel
    });

    try {
        if (!queue.connection) await queue.connect(interaction.member.voice.channel);
    } catch {
        void player.deleteQueue(interaction.guildId);
        return void interaction.followUp({ content: "Could not join your voice channel!" });
    }

    await interaction.followUp({ content: `⏱ | Loading your ${searchResult.playlist ? "playlist" : "track"}...` });
    searchResult.playlist ? queue.addTracks(searchResult.tracks) : queue.addTrack(searchResult.tracks[0]);
    if (!queue.playing) await queue.play();
}

Lastly, if the queue isn’t already playing, let’s start it using the play() function.

How to skip songs

Skipping is quite easy – you can do it by calling the skip() function on the queue.

if (interaction.commandName === "skip") {
    await interaction.deferReply();
    const queue = player.getQueue(interaction.guildId);
    if (!queue || !queue.playing) return void interaction.followUp({ content: "❌ | No music is being played!" });
    const currentTrack = queue.current;
    const success = queue.skip();
    return void interaction.followUp({
        content: success ? `✅ | Skipped **${currentTrack}**!` : "❌ | Something went wrong!"
    });
}

If the action is successful, you can write a message to the Discord text channel using interaction.followUp().

How to stop songs

The stop functionality will remove all the songs from the queue and the bot will leave the voice channel. You can do this by destroying the current queue which automatically makes the bot leave the voice channel (unless you configure it otherwise in the player configuration).

else if (interaction.commandName === "stop") {
        await interaction.deferReply();
        const queue = player.getQueue(interaction.guildId);
        if (!queue || !queue.playing) return void interaction.followUp({ content: "❌ | No music is being played!" });
        queue.destroy();
        return void interaction.followUp({ content: "🛑 | Stopped the player!" });
    }

Complete source code for the index.js:

Here you can get the complete source code for the music bot:

const { Client, GuildMember, Intents } = require("discord.js");
const { Player, QueryType } = require("discord-player");
const config = require("./config.json");

const client = new Client({
    intents: [Intents.FLAGS.GUILD_VOICE_STATES, Intents.FLAGS.GUILD_MESSAGES, Intents.FLAGS.GUILDS]
});

client.on("ready", () => {
    console.log("Bot is online!");
    client.user.setActivity({
        name: "🎶 | Music Time",
        type: "LISTENING"
    });
});
client.on("error", console.error);
client.on("warn", console.warn);

const player = new Player(client);

player.on("error", (queue, error) => {
    console.log(`[${queue.guild.name}] Error emitted from the queue: ${error.message}`);
});
player.on("connectionError", (queue, error) => {
    console.log(`[${queue.guild.name}] Error emitted from the connection: ${error.message}`);
});

player.on("trackStart", (queue, track) => {
    queue.metadata.send(`🎶 | Started playing: **${track.title}** in **${queue.connection.channel.name}**!`);
});

player.on("trackAdd", (queue, track) => {
    queue.metadata.send(`🎶 | Track **${track.title}** queued!`);
});

player.on("botDisconnect", (queue) => {
    queue.metadata.send("❌ | I was manually disconnected from the voice channel, clearing queue!");
});

player.on("channelEmpty", (queue) => {
    queue.metadata.send("❌ | Nobody is in the voice channel, leaving...");
});

player.on("queueEnd", (queue) => {
    queue.metadata.send("✅ | Queue finished!");
});

client.on("messageCreate", async (message) => {
    if (message.author.bot || !message.guild) return;
    if (!client.application?.owner) await client.application?.fetch();

    if (message.content === "!deploy" && message.author.id === client.application?.owner?.id) {
        await message.guild.commands.set([
            {
                name: "play",
                description: "Plays a song from youtube",
                options: [
                    {
                        name: "query",
                        type: "STRING",
                        description: "The song you want to play",
                        required: true
                    }
                ]
            },
            {
                name: "skip",
                description: "Skip to the current song"
            },
            {
                name: "stop",
                description: "Stop the player"
            },
        ]);

        await message.reply("Deployed!");
    }
});

client.on("interactionCreate", async (interaction) => {
    if (!interaction.isCommand() || !interaction.guildId) return;

    if (!(interaction.member instanceof GuildMember) || !interaction.member.voice.channel) {
        return void interaction.reply({ content: "You are not in a voice channel!", ephemeral: true });
    }

    if (interaction.guild.me.voice.channelId && interaction.member.voice.channelId !== interaction.guild.me.voice.channelId) {
        return void interaction.reply({ content: "You are not in my voice channel!", ephemeral: true });
    }

    if (interaction.commandName === "play") {
        await interaction.deferReply();

        const query = interaction.options.get("query").value;
        const searchResult = await player
            .search(query, {
                requestedBy: interaction.user,
                searchEngine: QueryType.AUTO
            })
            .catch(() => {});
        if (!searchResult || !searchResult.tracks.length) return void interaction.followUp({ content: "No results were found!" });

        const queue = await player.createQueue(interaction.guild, {
            metadata: interaction.channel
        });

        try {
            if (!queue.connection) await queue.connect(interaction.member.voice.channel);
        } catch {
            void player.deleteQueue(interaction.guildId);
            return void interaction.followUp({ content: "Could not join your voice channel!" });
        }

        await interaction.followUp({ content: `⏱ | Loading your ${searchResult.playlist ? "playlist" : "track"}...` });
        searchResult.playlist ? queue.addTracks(searchResult.tracks) : queue.addTrack(searchResult.tracks[0]);
        if (!queue.playing) await queue.play();
    } else if (interaction.commandName === "skip") {
        await interaction.deferReply();
        const queue = player.getQueue(interaction.guildId);
        if (!queue || !queue.playing) return void interaction.followUp({ content: "❌ | No music is being played!" });
        const currentTrack = queue.current;
        const success = queue.skip();
        return void interaction.followUp({
            content: success ? `✅ | Skipped **${currentTrack}**!` : "❌ | Something went wrong!"
        });
    } else if (interaction.commandName === "stop") {
        await interaction.deferReply();
        const queue = player.getQueue(interaction.guildId);
        if (!queue || !queue.playing) return void interaction.followUp({ content: "❌ | No music is being played!" });
        queue.destroy();
        return void interaction.followUp({ content: "🛑 | Stopped the player!" });
    } else {
        interaction.reply({
            content: "Unknown command!",
            ephemeral: true
        });
    }
});

client.login(config.token);

Discord Bot Version 0.12

Now you'll just need to install some dependencies before we can get started.

npm install discord.js ffmpeg fluent-ffmpeg @discordjs/opus ytdl-core --save

After installing the dependencies, import them in your dependencies.

const Discord = require('discord.js');
const {
    prefix,
    token,
} = require('./config.json');
const ytdl = require('ytdl-core');

After that, create your client and login using your token.

const client = new Discord.Client();
client.login(token);

Now let’s add some basic listeners that console.log when they get executed.

client.once('ready', () => {
 console.log('Ready!');
});
client.once('reconnecting', () => {
 console.log('Reconnecting!');
});
client.once('disconnect', () => {
 console.log('Disconnect!');
});

After that, you can start your bot using the node command and it should be online on Discord and print “Ready!” in the console.

node index.js

How to read messages

Now that your bot is on your server and able to go online, you can start reading chat messages and responding to them.

To read messages, you only need to write one simple function:

client.on('message', async message => {

}

Here, you're creating a listener for the message event, getting the message, and saving it into a message object if it's triggered.

Now you need to check if the message is from your own bot and ignore it if it is.

if (message.author.bot) return;

In this line, you're checking if the author of the message is your bot and returning if it is.

After that, check if the message starts with the prefix you defined earlier and return if it doesn’t.

if (!message.content.startsWith(prefix)) return;

After that, you can check which command you need to execute. You can do so using some simple if statements:

const serverQueue = queue.get(message.guild.id);

if (message.content.startsWith(`${prefix}play`)) {
    execute(message, serverQueue);
    return;
} else if (message.content.startsWith(`${prefix}skip`)) {
    skip(message, serverQueue);
    return;
} else if (message.content.startsWith(`${prefix}stop`)) {
    stop(message, serverQueue);
    return;
} else {
    message.channel.send("You need to enter a valid command!");
}

In this code block, you're checking which command to execute and calling the command. If the input command isn’t valid, you're writing an error message into the chat using the send() function.

Now that you know which command you need to execute, you can start implementing these commands.

How to add songs

Let’s start by adding the play command. For that, you'll need a song and a guild (a guild represents an isolated collection of users and channels and is often referred to as a server). You'll also need the ytdl library you installed earlier.

First, create a map with the name of the queue where you save all the songs you type in the chat.

const queue = new Map();

After that, create an async function called execute and check if the user is in a voice chat and if the bot has the right permissions. If not, write an error message and return.

async function execute(message, serverQueue) {
  const args = message.content.split(" ");

  const voiceChannel = message.member.voice.channel;
  if (!voiceChannel)
    return message.channel.send(
      "You need to be in a voice channel to play music!"
    );
  const permissions = voiceChannel.permissionsFor(message.client.user);
  if (!permissions.has("CONNECT") || !permissions.has("SPEAK")) {
    return message.channel.send(
      "I need the permissions to join and speak in your voice channel!"
    );
  }
}

Now you can continue with getting the song info and saving it into a song object. For that, use your ytdl library which gets the song information from the YouTube link.

const songInfo = await ytdl.getInfo(args[1]);
const song = {
 title: songInfo.title,
 url: songInfo.video_url,
};

This will get the information of the song using the ytdl library you installed earlier. Then, save the information you need into a song object.

After saving the song info, you just need to create a contract you can add to your queue.

To do so, first check if your serverQueue is already defined which means that music is already playing. If so, add the song to your existing serverQueue and send a success message. If not, create it and try to join the voice channel and start playing music.

if (!serverQueue) {

}else {
 serverQueue.songs.push(song);
 console.log(serverQueue.songs);
 return message.channel.send(`${song.title} has been added to the queue!`);
}

Here, check if the serverQueueis empty and add the song to it if it’s not. Now you just need to create your contract if the serverQueue is null.

// Creating the contract for our queue
const queueContruct = {
 textChannel: message.channel,
 voiceChannel: voiceChannel,
 connection: null,
 songs: [],
 volume: 5,
 playing: true,
};
// Setting the queue using our contract
queue.set(message.guild.id, queueContruct);
// Pushing the song to our songs array
queueContruct.songs.push(song);

try {
 // Here we try to join the voicechat and save our connection into our object.
 var connection = await voiceChannel.join();
 queueContruct.connection = connection;
 // Calling the play function to start a song
 play(message.guild, queueContruct.songs[0]);
} catch (err) {
 // Printing the error message if the bot fails to join the voicechat
 console.log(err);
 queue.delete(message.guild.id);
 return message.channel.send(err);
}

In this code block, you created a contract and added your song to the songs array. After that, you tried to join the voice chat of the user and called your play() function you'll implement after that.

How to play songs

Now that you can add our songs to your queue and create a contract if there isn’t one yet, you can implement the play functionality.

First, create a function called play which takes two parameters (the guild and the song you want to play) and checks if the song is empty. If so, just leave the voice channel and delete the queue.

function play(guild, song) {
  const serverQueue = queue.get(guild.id);
  if (!song) {
    serverQueue.voiceChannel.leave();
    queue.delete(guild.id);
    return;
  }
}

After that, start playing your song using the play() function of the connection and passing the URL of your song.

const dispatcher = serverQueue.connection
    .play(ytdl(song.url))
    .on("finish", () => {
        serverQueue.songs.shift();
        play(guild, serverQueue.songs[0]);
    })
    .on("error", error => console.error(error));
dispatcher.setVolumeLogarithmic(serverQueue.volume / 5);
serverQueue.textChannel.send(`Start playing: **${song.title}**`);

Here, you created a stream and passed it the URL of our song. You also added two listeners that handle the end and error events.

Note: This is a recursive function which means that it calls itself over and over again. We're using recursion so it plays the next song when the song is finished.

Now you're ready to play a song by just typing the !play URL in the chat.

How to skip songs

Now you can implement the skipping functionality. For that, you just need to end the dispatcher you created in your play() function so it starts the next song.

function skip(message, serverQueue) {
  if (!message.member.voice.channel)
    return message.channel.send(
      "You have to be in a voice channel to stop the music!"
    );
  if (!serverQueue)
    return message.channel.send("There is no song that I could skip!");
  serverQueue.connection.dispatcher.end();
}

Here, you're checking if the user that typed the command is in a voice channel and if there is a song to skip.

How to stop songs

The stop() function is almost the same as skip(), except that you clear the songs array which will make your bot delete the queue and leave the voice chat.

function stop(message, serverQueue) {
  if (!message.member.voice.channel)
    return message.channel.send(
      "You have to be in a voice channel to stop the music!"
    );
  serverQueue.songs = [];
  serverQueue.connection.dispatcher.end();
}

Complete source code for the index.js:

Here you can get the complete source code for the music bot:

const Discord = require("discord.js");
const { prefix, token } = require("./config.json");
const ytdl = require("ytdl-core");

const client = new Discord.Client();

const queue = new Map();

client.once("ready", () => {
  console.log("Ready!");
});

client.once("reconnecting", () => {
  console.log("Reconnecting!");
});

client.once("disconnect", () => {
  console.log("Disconnect!");
});

client.on("message", async message => {
  if (message.author.bot) return;
  if (!message.content.startsWith(prefix)) return;

  const serverQueue = queue.get(message.guild.id);

  if (message.content.startsWith(`${prefix}play`)) {
    execute(message, serverQueue);
    return;
  } else if (message.content.startsWith(`${prefix}skip`)) {
    skip(message, serverQueue);
    return;
  } else if (message.content.startsWith(`${prefix}stop`)) {
    stop(message, serverQueue);
    return;
  } else {
    message.channel.send("You need to enter a valid command!");
  }
});

async function execute(message, serverQueue) {
  const args = message.content.split(" ");

  const voiceChannel = message.member.voice.channel;
  if (!voiceChannel)
    return message.channel.send(
      "You need to be in a voice channel to play music!"
    );
  const permissions = voiceChannel.permissionsFor(message.client.user);
  if (!permissions.has("CONNECT") || !permissions.has("SPEAK")) {
    return message.channel.send(
      "I need the permissions to join and speak in your voice channel!"
    );
  }

  const songInfo = await ytdl.getInfo(args[1]);
  const song = {
    title: songInfo.title,
    url: songInfo.video_url
  };

  if (!serverQueue) {
    const queueContruct = {
      textChannel: message.channel,
      voiceChannel: voiceChannel,
      connection: null,
      songs: [],
      volume: 5,
      playing: true
    };

    queue.set(message.guild.id, queueContruct);

    queueContruct.songs.push(song);

    try {
      var connection = await voiceChannel.join();
      queueContruct.connection = connection;
      play(message.guild, queueContruct.songs[0]);
    } catch (err) {
      console.log(err);
      queue.delete(message.guild.id);
      return message.channel.send(err);
    }
  } else {
    serverQueue.songs.push(song);
    return message.channel.send(`${song.title} has been added to the queue!`);
  }
}

function skip(message, serverQueue) {
  if (!message.member.voice.channel)
    return message.channel.send(
      "You have to be in a voice channel to stop the music!"
    );
  if (!serverQueue)
    return message.channel.send("There is no song that I could skip!");
  serverQueue.connection.dispatcher.end();
}

function stop(message, serverQueue) {
  if (!message.member.voice.channel)
    return message.channel.send(
      "You have to be in a voice channel to stop the music!"
    );
  serverQueue.songs = [];
  serverQueue.connection.dispatcher.end();
}

function play(guild, song) {
  const serverQueue = queue.get(guild.id);
  if (!song) {
    serverQueue.voiceChannel.leave();
    queue.delete(guild.id);
    return;
  }

  const dispatcher = serverQueue.connection
    .play(ytdl(song.url))
    .on("finish", () => {
      serverQueue.songs.shift();
      play(guild, serverQueue.songs[0]);
    })
    .on("error", error => console.error(error));
  dispatcher.setVolumeLogarithmic(serverQueue.volume / 5);
  serverQueue.textChannel.send(`Start playing: **${song.title}**`);
}

client.login(token);

Conclusion

You made it all the way until the end! Hope that this article helped you understand the Discord API and how you can use it to create a simple bot.

If you want to see an example of a more advanced discord bot, you can visit my GitHub repository.

If you have found this useful, please consider recommending and sharing it with other fellow developers.

If you have any questions or feedback, let me know and I'd be happy to help.

This tutorial discusses everything you need to know to transform HTML elements like a pro.

Table of Contents

1. What is the CSS transform Property?

2. What is the CSS rotate() Function?

3. What is the CSS rotateX() Function?

4. What is the CSS rotateY() Function?

5. What is the CSS rotateZ() Function?

6. What is the CSS rotate3d() Function?

7. CSS Rotate Functions vs. rotate Property: What's the Difference?

8. What is the CSS scale() Function?

9. CSS scale() Function vs. scale Property: What's the Difference?

10. What is the CSS scaleZ() Function?

11. What is the CSS scale3d() Function?

12. What is the CSS skew() Function?

13. What is the CSS translate() Function?

14. What is the CSS translateZ() Function?

15. What is the CSS translate3d() Function?

16. CSS Translate Functions vs. translate Property: What's the Difference?

17. What is the CSS perspective() Function?

18. CSS perspective() Function vs. perspective Property: What's the Difference?

19. What is the CSS matrix() Function?

20. Why Does the CSS Transform Functions' Order Matter?

21. Tools for Converting Transform Functions to matrix()

22. Important Stuff to Know about Transforming Elements in CSS

23. Wrapping Up

So, without further ado, let's discuss the CSS transform property.

What Is the CSS transform Property?

The CSS transform property specifies the transformational effect you wish to apply to an HTML element.

Here's the syntax:

html-element {
  transform: value;
}

The CSS transform property accepts the following values:

• inherit: Transforms the element with its parent element's transform value.

• initial: Transforms the HTML element with its default transform value.

• matrix(): Transforms the element two-dimensionally with a matrix of six values.

• matrix3d(): Transforms the HTML element three-dimensionally with a 4x4 matrix of sixteen values.

• none: Applies no transformation to the HTML element.

• perspective(): Transforms a 3D transformed element with a perspective view.

• rotate(): Transforms the element by rotating it two-dimensionally.

• rotate3d(): Transforms the element by rotating it three-dimensionally.

• rotateX(): Transforms the element by rotating it three-dimensionally along the X-axis.

• rotateY(): Transforms the element by rotating it three-dimensionally along the Y-axis.

• rotateZ(): Transforms the HTML element by rotating it three-dimensionally along the Z-axis.

• scale(): Transforms the element by scaling it two-dimensionally.

• scale3d(): Transforms the element by scaling it three-dimensionally.

• scaleX(): Transforms the element by scaling it along the X-axis.

• scaleY(): Transforms the element by scaling it along the Y-axis.

• scaleZ(): Transforms the HTML element by scaling it three-dimensionally along the Z-axis.

• skew(): Transforms the element by skewing it two-dimensionally along the X- and Y-axis.

• skewX(): Transforms the element by skewing it two-dimensionally along the X-axis.

• skewY(): Transforms the element by skewing it two-dimensionally along the Y-axis.

• translate(): Transforms the HTML element by translating (moving) it two-dimensionally.

• translate3d(): Transforms the element by translating it three-dimensionally.

• translateX(): Transforms the element by translating it along the X-axis.

• translateY(): Transforms the element by translating it along the Y-axis.

• translateZ(): Transforms the element by translating it three-dimensionally along the Z-axis.

Note: The transform property accepts one or more CSS transform functions. For instance, here's a valid transform declaration:

div {
  transform: perspective(370px) scaleZ(5) rotate(17deg);
}

In the snippet above, we assigned three transform functions to the transform property. Let's talk more about some of transform's values.

What is the CSS rotate() Function?

rotate() transforms an element by rotating it two-dimensionally around a fixed point.

Note:

• "Transform origin" is the fixed point around which an element rotates.

• You can define your element's fixed point using the CSS transform-origin property. But the default is center.

Syntax of the CSS rotate() function

rotate() accepts a single argument. Here is the syntax:

element {
  transform: rotate(angle);
}

Note the following:

• The rotate(angle) function is equivalent to rotate3d(0, 0, 1, angle) or rotateZ(angle).

• The angle argument specifies the element's angle of rotation.

• angle can be in degrees, gradians, radians, or turns.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

• Your browser's writing direction determines the element's direction of rotation.

• A positive angle will rotate the element clockwise in a left-to-right writing direction. But a negative angle will do a counterclockwise rotation.

• A positive angle will rotate the element counterclockwise in a right-to-left writing context. But a negative angle will do a clockwise rotation.

Examples of the CSS rotate() function

Below are some examples of how the CSS rotate() function works.

How to do a zero-degree rotation in CSS:

img {
  transform: rotate(0deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate() function to specify a zero-degree (0⁰) rotation for the image element.

How to do a 45-degree rotation in CSS:

img {
  transform: rotate(45deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate() function to specify a forty-five-degree (45⁰) rotation for the image element.

How to do a negative seventy-degree rotation in CSS:

img {
  transform: rotate(-70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate() function to specify a negative seventy-degree (70⁰) rotation for the image element.

What is the CSS rotateX() Function?

rotateX() transforms an element by rotating it three-dimensionally around the X-axis.

A three-dimensional Cartesian coordinate system showing the X-, Y-, and Z-axis

Syntax of the CSS rotateX() function

rotateX() accepts a single argument. Here is the syntax:

element {
  transform: rotateX(angle);
}

Note the following:

• The rotateX(angle) function is equivalent to rotate3d(1, 0, 0, angle).

• The angle argument specifies the element's angle of rotation.

• angle can be in degree, gradian, radian, or turn.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

Examples of the CSS rotateX() function

Below are some examples of how the CSS rotateX() function works.

How to do a zero-degree rotation around the X-axis:

img {
  transform: rotateX(0deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateX() function to specify a zero-degree (0⁰) rotation for the image around the X-axis.

How to do a 70-degree rotation around the X-axis:

img {
  transform: rotateX(70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateX() function to specify a seventy-degree (70⁰) rotation for the image around the X-axis.

What is the CSS rotateY() Function?

rotateY() transforms an element by rotating it three-dimensionally around the Y-axis.

Syntax of the CSS rotateY() function

rotateY() accepts a single argument. Here is the syntax:

element {
  transform: rotateY(angle);
}

Note the following:

• The rotateY(angle) function is equivalent to rotate3d(0, 1, 0, angle).

• The angle argument specifies the element's angle of rotation.

• angle can be in degrees, gradians, radians, or turns.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

Examples of the CSS rotateY() function

Below are some examples of how the CSS rotateY() function works.

How to do a zero-degree rotation around the Y-axis:

img {
  transform: rotateY(0deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateY() function to specify a zero-degree (0⁰) rotation for the image around the Y-axis.

How to do a 70-degree rotation around the Y-axis:

img {
  transform: rotateY(70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateY() function to specify a seventy-degree (70⁰) rotation for the image around the Y-axis.

What is the CSS rotateZ() Function?

rotateZ() transforms an element by rotating it three-dimensionally around the Z-axis.

Syntax of the CSS rotateZ() function

rotateZ() accepts a single argument. Here is the syntax:

element {
  transform: rotateZ(angle);
}

Note the following:

• The rotateZ(angle) function is equivalent to rotate3d(0, 0, 1, angle) or rotate(angle).

• The angle argument specifies the element's angle of rotation.

• angle can be in degrees, gradians, radians, or turns.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

Examples of the CSS rotateZ() function

Below are some examples of how the CSS rotateZ() function works.

How to do a zero-degree rotation around the Z-axis:

img {
  transform: rotateZ(0deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateZ() function to specify a zero-degree (0⁰) rotation for the image around the Z-axis.

How to do a 70-degree rotation around the Z-axis:

img {
  transform: rotateZ(70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotateZ() function to specify a seventy-degree (70⁰) rotation for the image around the Z-axis.

What is the CSS rotate3d() Function?

rotate3d() transforms an element by rotating it three-dimensionally around the x-, y-, and z-axis.

Syntax of the CSS rotate3d() function

rotate3d() accepts four arguments. Here is the syntax:

element {
  transform: rotate3d(x, y, z, angle);
}

Note the following:

• The x, y, and z arguments are numbers specifying the x-, y-, and z-coordinates.

• The coordinates are the axis around which the element will rotate.

• The angle argument specifies the element's angle of rotation.

• angle can be in degrees, gradians, radians, or turns.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

Examples of the CSS rotate3d() function

Below are some examples of how the CSS rotate3d() function works.

How to do a 70-degree rotation around the Z-axis:

img {
  transform: rotate3d(0, 0, 1, 70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate3d() function to specify a seventy-degree (70⁰) rotation for the image around the Z-axis.

How to do a 70-degree rotation around the X-, Y-, and Z-axis:

img {
  transform: rotate3d(1, 1, 1, 70deg);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate3d() function to specify a seventy-degree (70⁰) rotation for the image around the x-, y-, and z-axis.

CSS Rotate Functions vs. rotate Property: What's the Difference?

CSS rotate functions and CSS rotate property provides two similar ways to specify rotation transformations.

The main differences between the two rotation techniques are as follows:

• The CSS rotate property allows rotating an element without using the CSS transform property.

• The CSS rotate property's syntax is shorter than its function alternative.

• The CSS rotate property saves you from remembering the specific order to position the transform functions.

• Browsers calculate the transform functions' matrix in the order you assigned them to the CSS transform property—from left to right.

• Browsers calculate the transform properties' matrix in the following transformation matrix order:

  1. translate

  2. rotate

  3. scale

Below are some examples.

How to use CSS rotate property vs. function to do a 45-degree rotation

img {
  rotate: 45deg; /* Equivalent to a transform: rotate(45deg) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate property to specify a forty-five-degree (45⁰) rotation for the image element.

How to use CSS rotate property vs. function to do a 70-degree rotation around the X-axis

img {
  rotate: x 70deg; /* Equal to a transform: rotateX(70deg) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate property to specify a seventy-degree (70⁰) rotation for the image around the X-axis.

How to use CSS rotate property vs. function to do a 70-degree rotation around the Y-axis

img {
  rotate: y 70deg; /* Equal to a transform: rotateY(70deg) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate property to specify a seventy-degree (70⁰) rotation for the image around the Y-axis.

How to use CSS rotate property vs. function to do a 70-degree rotation around the Z-axis

img {
  rotate: z 70deg; /* Equal to a transform: rotateZ(70deg) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate property to specify a seventy-degree (70⁰) rotation for the image around the Z-axis.

How to use CSS rotate property vs. function to do a 70-degree rotation around the X-, Y-, and Z-axis

img {
  rotate: 1 1 1 70deg; /* Equal to a transform: rotate3d(1, 1, 1, 70deg) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the rotate property to specify a seventy-degree (70⁰) rotation for the image around the x-, y-, and z-axis.

Note: A none value tells browsers not to rotate the selected element.

What is the CSS scale() Function?

scale() transforms an element by resizing (scaling) it two-dimensionally from a fixed point.

Note:

• "Transform origin" is the fixed point from which the computer scales an element.

• You can define your element's fixed point using the CSS transform-origin property. But the default is center.

Syntax of the CSS scale() function

scale() accepts two arguments. Here is the syntax:

element {
  transform: scale(x, y);
}

Note the following:

• The x argument can be a number or percentage. It specifies the element's scaling factor along the x-axis.

• The y argument can also be a number or percentage. It defines the element's scaling factor along the y-axis.

• Y-axis' default value is x. Therefore, if you do not provide a y argument, the browser automatically uses x's value.

• Suppose x and y are equal. In that case, browsers will scale your element uniformly and preserve its aspect ratio.

A two-dimensional Cartesian coordinate system showing the X- and Y-axis

Examples of the CSS scale() function

Below are some examples of how the CSS scale() function works.

How to scale an element uniformly along the X- and Y-axis in CSS:

img {
  transform: scale(0.3);
  transform-origin: left;
}

Try it on StackBlitz

The snippet above used the scale() function to specify a 0.3 scaling factor for the image element along the X- and Y-axis.

Note:

• scale(0.3) is equivalent to scale(0.3, 0.3).

• The percentage equivalence of scale(0.3) is scale(30%).

How to scale an element non-uniformly along the X- and Y-axis in CSS:

img {
  transform: scale(0.3, 65%);
  transform-origin: top left;
}

Try it on StackBlitz

The snippet above used the scale() function to specify a 0.3 scaling factor for the image along the X-axis and 65% along the Y-axis.

How to scale an element along only the X-axis:

img {
  transform: scale(0.3, 1);
  transform-origin: top left;
}

Try it on StackBlitz

The snippet above used the scale() function to specify a 0.3 scaling factor for the image along only the X-axis.

Note:

• A scale factor of 1 or 100% tells browsers not to apply any scaling effect on the selected element.

• scale(0.3, 1) is equivalent to scaleX(0.3).

How to scale an element along only the Y-axis:

img {
  transform: scale(100%, 0.2);
  transform-origin: top left;
}

Try it on StackBlitz

The snippet above used the scale() function to specify a 0.2 scaling factor for the image along only the Y-axis.

Note:

• A 100% or 1 scale factor tells browsers not to apply any scaling effect on the selected element.

• scale(100%, 0.2) is equivalent to scaleY(0.2).

CSS scale() Function vs. scale Property: What's the Difference?

The CSS scale() function and the CSS scale property provide two similar ways to specify a scale transformation.

The main differences between the two scaling techniques are as follows:

• The CSS scale property allows scaling an element without using the CSS transform property.

• The CSS scale property's syntax is shorter than its function alternative.

• The CSS scale property saves you from remembering the specific order to position the transform functions.

• Browsers calculate the transform functions' matrix in the order you assigned them to the CSS transform property—from left to right.

• Browsers calculate the transform properties' matrix in the following order:

  1. translate

  2. rotate

  3. scale

Here's an example:

Use the CSS scale property to scale an element nonuniformly along the X- and Y-axis.

img {
  scale: 0.3 65%; /* Equal to a transform: scale(0.3, 65%) property */
  transform-origin: top left;
}

Try it on StackBlitz

The snippet above used the scale property to specify a 0.3 scaling factor for the image along the X-axis and 65% along the Y-axis.

Note: A none value tells browsers not to scale the selected element.

What is the CSS scaleZ() Function?

scaleZ() transforms an element by resizing (scaling) it three-dimensionally from a fixed point along the z-axis.

Note:

• "Transform origin" is the fixed point from which the computer scales an element.

• You can define your element's fixed point using the CSS transform-origin property. But the default is center.

Syntax of the CSS scaleZ() function

scaleZ() accepts a single argument. Here is the syntax:

element {
  transform: scaleZ(number);
}

Note:

• The scaleZ(number) function is equivalent to scale3d(1, 1, number).

• The number argument specifies the element's scaling factor along the z-axis.

Examples of the CSS scaleZ() function

We often use scaleZ() with other CSS functions such as perspective(), translateZ(), and rotateX(). Below are some examples.

How to use scaleZ() with CSS perspective() and rotateX() functions:

img {
  transform: perspective(370px) scaleZ(5) rotateX(17deg);
  width: 80%;
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 370px distance between the user and the z=0 plane.

2. The scaleZ() function specifies a scale factor of 5 for the image along the z-axis.

3. We used the rotateX() function to rotate the image seventeen-degree (17⁰) around the x-axis.

Note:

• List perspective() first whenever you chain it with other CSS transform functions. Otherwise, browsers might transform the selected element incorrectly.

• List the scaleZ() function before rotateX(). Otherwise, the browser will not scale the element.

How to use scaleZ() with CSS perspective() and translateZ() functions:

img {
  width: 40%;
}

.second-image {
  transform: perspective(370px) scaleZ(5) translateZ(30px);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 370px distance between the user and the z=0 plane.

2. The scaleZ() function specifies a scale factor of 5 for the image along the z-axis.

3. We used the translateZ() function to reposition the second-image thirty pixels (30px) away from its original position along the z-axis.

What is the CSS scale3d() Function?

scale3d() transforms an element by resizing (scaling) it three-dimensionally from a fixed point along the x-, y-, and z-axis.

Note:

• "Transform origin" is the fixed point from which the computer scales an element.

• You can define your element's fixed point using the CSS transform-origin property. But the default is center.

Syntax of the CSS scale3d() function

scale3d() accepts three arguments. Here is the syntax:

element {
  transform: scale3d(x, y, z);
}

The x, y, and z arguments are numbers specifying the x-, y-, and z-coordinates. The coordinates are the axis along which browsers will scale the element.

Examples of the CSS scale3d() function

Below are some examples of how the CSS scale3d() function works.

How to use scale3d() with CSS perspective() and rotateX() functions:

img {
  transform: perspective(370px) scale3d(1, 1, 5) rotateX(17deg);
  width: 80%;
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 370px distance between the user and the z=0 plane.

2. The scale3d() function specifies a scale factor of 1, 1, and 5 for the image along the x-, y-, and z-axis.

3. We used the rotateX() function to rotate the image seventeen-degrees (17⁰) around the x-axis.

Note:

• A scale factor of 1 will apply no scaling effect on the element.

• List perspective() first whenever you chain it with other CSS transform functions. Otherwise, browsers might transform the selected element incorrectly.

• List the scale3d() function before rotateX(). Otherwise, the browser will not scale the element.

How to scale elements three-dimensionally:

img {
  width: 40%;
}

.second-image {
  transform: scale3d(5, 3, 0.05);
  transform-origin: top left;
}

Try it on StackBlitz

We used the scale3d() function to specify a scale factor of 5, 3, and 0.05 for the image along the x-, y-, and z-axis.

What is the CSS skew() Function?

skew() transforms an element by slanting (skewing) it two-dimensionally around a fixed point.

Note:

• "Transform origin" is the fixed point from which the computer skews an element.

• You can define your element's fixed point using the CSS transform-origin property. But the default is center.

Syntax of the CSS skew() function

skew() accepts two arguments. Here is the syntax:

element {
  transform: skew(aX, aY);
}

Note the following:

• The aX argument specifies the element's skewing angle along the x-axis.

• The aY argument specifies the element's skewing angle along the y-axis.

• aX and aY can be in degrees, gradians, radians, or turns.

• An angle argument consists of a number followed by the unit you wish to use—for instance, 45deg.

• aY is an optional argument.

Examples of the CSS skew() function

Below are some examples of how the CSS skew() function works.

How to skew an element along only the X-axis:

img {
  transform: skew(30deg);
  transform-origin: top;
  width: 80%;
}

Try it on StackBlitz

The snippet above used the skew() function to apply a thirty-degree (30⁰) slant on the image along only the x-axis.

Note: skew(30deg) is equivalent to skewX(30deg).

How to skew an element along only the Y-axis:

img {
  transform: skew(0, 40deg);
  transform-origin: top left;
  width: 80%;
}

Try it on StackBlitz

The snippet above used the skew() function to apply a forty-degree (40⁰) slant on the image along only the y-axis.

Note:

• A zero (0) skew degree tells browsers not to apply any skewing effect on the selected element.

• skew(0, 40deg) is equivalent to skewY(40deg).

How to skew an element along the X- and Y-axis:

img {
  transform: skew(30deg, 40deg);
  transform-origin: top left;
  width: 80%;
}

Try it on StackBlitz

The snippet above used the skew() function to apply a thirty-degree (30⁰) slant on the image along the x-axis. And forty-degree (40⁰) along the y-axis.

What is the CSS translate() Function?

translate() transforms an element by repositioning (translating) it two-dimensionally.

Syntax of the CSS translate() function

translate() accepts two arguments. Here is the syntax:

element {
  transform: translate(x, y);
}

Note the following:

• The x argument can be a length or percentage value. It specifies the distance you wish to move the element from its original x-axis position.

• The y argument can be a length or percentage value. It defines the distance you wish to move the element from its original y-axis position.

• y is an optional argument.

Examples of the CSS translate() function

Below are some examples of how the CSS translate() function works.

How to translate an element along only the X-axis:

img {
  transform: translate(150px);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the translate() function to reposition the image 150px away from its original position along the x-axis.

You can also write it like this, specifying the X:

img {
  transform: translateX(150px);
  width: 80%

Note: translate(150px) is equivalent to translateX(150px).

How to translate an element along only the Y-axis:

img {
  transform: translate(0, 55%);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the translate() function to reposition the image 55% away from its original position along the y-axis.

Note:

• A zero (0) translate distance tells browsers not to apply any translating effect on the selected element.

• translate(0, 55%) is equivalent to translateY(55%).

How to translate an element along the X- and Y-axis:

img {
  transform: translate(60%, 300px);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the translate() function to reposition the image 60% away from its original position along the x-axis and 300px from its y-axis.

What is the CSS translateZ() Function?

translateZ() transforms an element by repositioning (translating) it three-dimensionally along the z-axis.

A three-dimensional Cartesian coordinate system showing the X-, Y-, and Z-axis

Syntax of the CSS translateZ() function

translateZ() accepts a single argument. Here is the syntax:

element {
  transform: translateZ(length);
}

The length argument specifies the distance you wish to move the element from its original z-axis position.

A three-dimensional Cartesian coordinate system with a red arrow defining the green plane's translateZ length

Examples of the CSS translateZ() function

We often use translateZ() with the perspective() function. Below are some examples.

How to use translateZ() with the CSS perspective() function:

img {
  width: 40%;
}

.second-image {
  transform: perspective(33px) translateZ(10px);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 33px distance between the user and the z=0 plane.

2. We used the translateZ() function to reposition the second-image ten pixels (10px) away from its original position along the z-axis.

Note:

• Suppose the second-image's z-axis position is larger than or equal to the perspective() function's argument. In that case, the image will disappear as though it is behind the user. In other words, the selected item disappears when the user is in the same position as the element (or when the element is behind the user).

• The larger the user's distance to the element's z-axis position, the less intensive the perspective effect will be, and vice-versa.

How to use translateZ() with a 70px perspective:

img {
  width: 40%;
}

.second-image {
  transform: perspective(70px) translateZ(10px);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 70px distance between the user and the z=0 plane.

2. We used the translateZ() function to reposition the second-image ten pixels (10px) away from its original position along the z-axis.

What is the CSS translate3d() Function?

translate3d() transforms an element by repositioning (translating) it three-dimensionally along the x-, y-, and z-axis.

Syntax of the CSS translate3d() function

translate3d() accepts three arguments. Here is the syntax:

element {
  transform: translate3d(x, y, z);
}

Note the following:

• The x argument can be a length or percentage value. It specifies the distance you wish to move the element from its original x-axis position.

• The y argument can be a length or percentage value. It defines the distance you wish to move the element from its original y-axis position.

• z can only be a length—not a percentage. It defines the distance you wish to move the element from its original z-axis position.

Examples of the CSS translate3d() function

Below are some examples of how the CSS translate3d() function works.

How to translate an element along only the X-axis

img {
  transform: translate3d(150px, 0, 0);
  width: 80%;
}

Try it on StackBlitz

The snippet above used the translate3d() function to reposition the image 150px away from its original position along the x-axis.

Note: translate3d(150px, 0, 0) is equivalent to translateX(150px).

How to translate elements three-dimensionally:

img {
  width: 40%;
}

.second-image {
  transform: perspective(300px) translate3d(15%, 45%, 200px);
}

Try it on StackBlitz

The snippet above used the translate3d() function to reposition the image 15% away from its original position along the x-axis, 45% from its y-axis, and 200px from its z-axis.

CSS Translate Functions vs. translate Property: What's the Difference?

The CSS translate functions and the CSS translate property provide two similar ways to specify a translation transformation.

The main differences between the two translation techniques are as follows:

• The CSS translate property allows translating an element without using the CSS transform property.

• The CSS translate property's syntax is shorter than its function alternative.

• The CSS translate property saves you from remembering the specific order to position the transform functions.

• Browsers calculate the transform functions' matrix in the order you assigned them to the CSS transform property—from left to right.

• Browsers calculate the transform properties' matrix in the following order:

  1. translate

  2. rotate

  3. scale

Below are some examples.

How to use CSS translate property vs. function to translate an element along the X- and Y-axis

img {
  translate: 60% 300px; /* Equal to a transform: translate(60%, 300px) property */
  width: 80%;
}

Try it on StackBlitz

The snippet above used the translate property to reposition the image 60% away from its original position along the x-axis. And 300px from its y-axis.

Note: Suppose you wish to translate an element along the z-axis. In that case, set a perspective property on the "parent element" of the element you want to translate. Otherwise, the element will not move along its z-axis.

How to use CSS translate property vs. function to translate an element along the Z-axis

img {
  width: 40%;
}

div {
  perspective: 35px;
}

.second-image {
  translate: 0px 0px 17px; /* Equal to a transform: translateZ(17px) property */
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective property to define a 35px distance between the user and the z=0 plane.

2. We used the translate property to reposition the second-image seventeen pixels (17px) away from its original position along the z-axis.

How to use CSS translate property vs. function to translate an element three-dimensionally

img {
  width: 40%;
}

div {
  perspective: 300px;
}

.second-image {
  translate: 50% 25% 200px; /* Equal to a transform: translate3d(50%, 25%, 200px) property */
}

Try it on StackBlitz

The snippet above used the translate property to reposition the image 50% away from its original position along the x-axis, 25% from its y-axis, and 200px from its z-axis.

Note: A none value tells browsers not to translate the selected element.

What is the CSS perspective() Function?

perspective() transforms an element by adding some perspective effects to it.

Syntax of the CSS perspective() function

perspective() accepts only one argument. Here is the syntax:

element {
  transform: perspective(length);
}

The length argument specifies the user's distance to the z=0 plane.

A three dimensional Cartesian coordinate system with a red arrow defining the distance between the user and the z=0 plane

Examples of the CSS perspective() function

We often use perspective() with other CSS functions such as translateZ(), rotateX(), and rotateY(). Below are some examples.

How to use perspective() with the CSS translateZ() function:

img {
  width: 40%;
}

.second-image {
  transform: perspective(33px) translateZ(10px);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 33px distance between the user and the z=0 plane.

2. We used the translateZ() function to reposition the second-image ten pixels (10px) away from its original position along the z-axis.

Note the following:

• Suppose the second-image's z-axis position is larger than or equal to the perspective() function's argument. In that case, the image will disappear as though it is behind the user. In other words, the selected item disappears when the user is in the same position as the element (or when the element is behind the user).

• The larger the user's distance to the element's z-axis position, the less intensive the perspective effect will be, and vice-versa.

How to use perspective() with the CSS rotateY() function:

img {
  width: 40%;
}

.second-image {
  transform: perspective(33px) rotateY(-10deg);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 33px distance between the user and the z=0 plane.

2. We used the rotateY() function to rotate the second-image negative ten-degree (-10⁰) around the y-axis.

How to use perspective() with the CSS rotateX() function:

img {
  width: 40%;
}

.second-image {
  transform: perspective(33px) rotateX(17deg);
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective() function to define a 33px distance between the user and the z=0 plane.

2. We used the rotateX() function to rotate the second-image seventeen-degree (17⁰) around the x-axis.

CSS perspective() Function vs. perspective Property: What's the Difference?

The CSS perspective() function and the perspective property provide two similar ways to add perspective effects to HTML elements.

The main differences between the two perspective techniques are as follows:

• We apply the perspective() function "directly on the element" we want to add some perspective effects to.

• We apply the perspective property "on the parent element" of the element we want to add some perspective effects to.

• The perspective() function works as a transform property's value.

• The CSS perspective property allows you to create perspective effects without using the CSS transform property.

Here's an example:

Use CSS perspective property to add perspective effect to a child element:

img {
  width: 40%;
}

div {
  perspective: 33px;
}

.second-image {
  rotate: x 17deg;
}

Try it on StackBlitz

Here's what we did in the snippet above:

1. We used the perspective property to define a 33px distance between the user and the z=0 plane.

2. We used the rotate property to rotate the second-image seventeen-degree (17⁰) around the x-axis.

Note:

• The CSS perspective property saves you from remembering the specific order to position the transform functions.

• A none value tells browsers not to add any perspective effect to the selected element's children.

What is the CSS matrix() Function?

The CSS matrix() function is a shorthand for the following 2D transform functions:

• scaleX()

• skewY()

• skewX()

• scaleY()

• translateX()

• translateY()

In other words, instead of writing:

img {
  transform-origin: 0 0;
  transform: translateX(100px) translateY(250px) scaleX(2) scaleY(0.9)
    skewX(10deg) skewY(35deg);
  width: 80%;
}

Try it on StackBlitz

You can alternatively use the matrix() function to shorten your code like so:

img {
  transform-origin: 0 0;
  transform: matrix(2.24693, 0.630187, 0.352654, 0.9, 100, 250);
  width: 80%;
}

Try it on StackBlitz

The CSS matrix() function's syntax

The matrix() function accepts six values. Here's the syntax:

matrix(scaleX(), skewY(), skewX(), scaleY(), translateX(), translateY())

You can represent the CSS matrix's values as homogeneous coordinates on ℝℙ² like so:

| scX skX tX | ← x-axis
| skY scY tY | ← y-axis
|  0   0   1 | ← constants

Note the following:

• scX and skX are numbers describing an element's scale and skew linear transformation on the x-axis.

• tX is a number representing an element's translation on the x-axis.

• skY and scY are numbers describing an element's skew and scale linear transformation on the y-axis.

• tY is a number representing an element's translation on the y-axis.

• 0, 0, 1 are constants.

• We do not pass the constants as arguments to the matrix() function because the computer implies them automatically.

Examples of the CSS matrix() function

Below are some examples of the CSS matrix() function.

How to convert scaleX() to matrix() function:

Consider the following transform property:

img {
  transform-origin: 0 0;
  transform: scaleX(2);
  width: 80%;
}

Try it on StackBlitz

Here is the matrix() equivalent of the above scaleX() function:

img {
  transform-origin: 0 0;
  transform: matrix(2, 0, 0, 1, 0, 0); /* scX, skY, skX, scY, tX, tY */
  width: 80%;
}

Try it on StackBlitz

Let's also represent the matrix's values as homogeneous coordinates on ℝℙ²:

| 2 0 0 | ← x-axis
| 0 1 0 | ← y-axis
| 0 0 1 | ← constants

Below is another example.

How to convert translateY() to matrix() function:

img {
  transform-origin: 0 0;
  transform: translateY(250px);
  width: 80%;
}

Try it on StackBlitz

Here is the matrix() equivalent of the above translateY() function:

img {
  transform-origin: 0 0;
  transform: matrix(1, 0, 0, 1, 0, 250); /* scX, skY, skX, scY, tX, tY */
  width: 80%;
}

Try it on StackBlitz

Let's also represent the matrix's values as homogeneous coordinates on ℝℙ²:

| 1 0 0   | ← x-axis
| 0 1 250 | ← y-axis
| 0 0 1   | ← constants

Below is a third example.

How to convert translateX() and scale() to matrix() function:

img {
  transform-origin: 0 0;
  transform: translateX(100px) scale(2);
  width: 80%;
}

Try it on StackBlitz

Here is the syntax for converting the above transform property's value to matrix():

matrix = (translateX's homogeneous coordinates) x (scale's homogeneous coordinates)

Let's begin the conversion by defining translateX(100px)'s homogeneous coordinates:

| 1 0 100 | ← x-axis
| 0 1 0   | ← y-axis
| 0 0 1   | ← constants

Let's also define scale(2)'s homogeneous coordinates:

| 2 0 0 | ← x-axis
| 0 2 0 | ← y-axis
| 0 0 1 | ← constants

It's now time to multiply the two homogeneous coordinates by using the following syntax:

| a d g |   | j m p |   | aj + dk + gl   am + dn + go   ap +dq  + gr |
| b e h | x | k n q | = | bj + ek + hl   bm + en + ho   bp + eq + hr |
| c f i |   | l o r |   | cj + fk + il   cm + fn + io   cp + fq + ir |

Let's implement the above syntax like so:

| 1 0 100 |   | 2 0 0 |   | 2 + 0 + 0   0 + 0 + 0   0 + 0 + 100 |
| 0 1  0  | x | 0 2 0 | = | 0 + 0 + 0   0 + 2 + 0   0 + 0 +  0  |
| 0 0  1  |   | 0 0 1 |   | 0 + 0 + 0   0 + 0 + 0   0 + 0 +  1  |

The next step is to resolve the addition. So, let's do that now.

| 1 0 100 |   | 2 0 0 |   | 2 0 100 |
| 0 1  0  | x | 0 2 0 | = | 0 2  0  |
| 0 0  1  |   | 0 0 1 |   | 0 0  1  |

The addition's result above gives us the homogeneous coordinates of the transform: translateX(100px) scale(2) property.

In other words, the product of (translateX's homogeneous coordinates) and (scale's homogeneous coordinates) equal:

| 2 0 100 | ← x-axis
| 0 2  0  | ← y-axis
| 0 0  1  | ← constants

Therefore, the matrix equivalence of transform: translateX(100px) scale(2) is transform: matrix(2, 0, 0, 2, 100, 0).

img {
  transform-origin: 0 0;
  transform: matrix(2, 0, 0, 2, 100, 0);
  width: 80%;
}

Try it on StackBlitz

Please note that transform: translateX(100px) scale(2) and transform: scale(2) translateX(100px) return different matrixes. Let's see an example of the second arrangement below.

How to convert scale() and translateX() to matrix() function:

Consider the following transform property:

img {
  transform-origin: 0 0;
  transform: scale(2) translateX(100px);
  width: 80%;
}

Try it on StackBlitz

Here is the syntax for converting the above transform property's value to matrix():

matrix = (scale's homogeneous coordinates) x (translateX's homogeneous coordinates)

Let's begin the conversion by defining scale(2)'s homogeneous coordinates:

| 2 0 0 | ← x-axis
| 0 2 0 | ← y-axis
| 0 0 1 | ← constants

Let's also define translateX(100px)'s homogeneous coordinates:

| 1 0 100 | ← x-axis
| 0 1  0  | ← y-axis
| 0 0  1  | ← constants

It's now time to multiply the two homogeneous coordinates by using the following syntax:

| a d g |   | j m p |   | aj + dk + gl   am + dn + go   ap +dq  + gr |
| b e h | x | k n q | = | bj + ek + hl   bm + en + ho   bp + eq + hr |
| c f i |   | l o r |   | cj + fk + il   cm + fn + io   cp + fq + ir |

Let's implement the above syntax like so:

| 2 0 0 |   | 1 0 100 |   | 2 + 0 + 0   0 + 0 + 0   200 + 0 + 0 |
| 0 2 0 | x | 0 1  0  | = | 0 + 0 + 0   0 + 2 + 0    0 + 0 + 0  |
| 0 0 1 |   | 0 0  1  |   | 0 + 0 + 0   0 + 0 + 0    0 + 0 + 1  |

The next step is to resolve the addition. So, let's do that now.

| 2 0 0 |   | 1 0 100 |   | 2 0 200 |
| 0 2 0 | x | 0 1  0  | = | 0 2  0  |
| 0 0 1 |   | 0 0  1  |   | 0 0  1  |

The addition's result above gives us the homogeneous coordinates of the transform: scale(2) translateX(100px) property.

In other words, the product of (scale's homogeneous coordinates) and (translateX's homogeneous coordinates) equal:

| 2 0 200 | ← x-axis
| 0 2  0  | ← y-axis
| 0 0  1  | ← constants

Therefore, the matrix equivalence of transform: scale(2) translateX(100px) is transform: matrix(2, 0, 0, 2, 200, 0).

img {
  transform-origin: 0 0;
  transform: matrix(2, 0, 0, 2, 200, 0);
  width: 80%;
}

Try it on StackBlitz

Notice that transform: scale(2) translateX(100px) equals transform: matrix(2, 0, 0, 2, 200, 0). And transform: translateX(100px) scale(2) is equivalent to transform: matrix(2, 0, 0, 2, 100, 0).

In other words, the order in which you write the transform functions matters. Let's discuss more on this below.

Why Does the CSS Transform Functions' Order Matter?

The order in which you write CSS transform functions matters because of the way browsers calculate the matrix's values.

For instance, consider the following snippet:

div {
  position: absolute;
  width: 100px;
  height: 100px;
  transform-origin: 0 0;
}

.red {
  border: 3px solid red;
  background-color: rgba(255, 0, 0, 0.5);
}

.green {
  border: 3px solid green;
  background-color: rgba(0, 128, 0, 0.5);
  transform: translateX(100px) scale(2);
}

.blue {
  border: 3px solid blue;
  background-color: rgba(0, 0, 255, 0.5);
  transform: scale(2) translateX(100px);
}

Try it on StackBlitz

The only difference between the green and the blue divs is the order in which we wrote their transform functions.

However, the computer translated the two containers using different values (100px for the green div and 200px for the blue one).

So, why did the transform functions' order affect the divs' translation values? Here's the reason:

• Browsers multiply each transform function's homogeneous coordinates in order—from left to right.

In other words, the computer used the following syntax to compute the green div's matrix:

• Green div's matrix = (translateX's homogeneous coordinates) x (scale's homogeneous coordinates)

And it used the following syntax to calculate the blue div's matrix:

• Blue div's matrix = (scale's homogeneous coordinates) x (translateX's homogeneous coordinates)

Therefore, the position of the transform functions determined the matrix's arguments because browsers began the calculation in order from the leftmost function to the right.

Knowing how to convert transform functions to matrix() is beneficial. And having some conversion tools can come in handy. So, let's discuss some helpful tools you can use.

Tools for Converting Transform Functions to matrix()

The two tools you can use to do a quick conversion of transform functions to matrix() are:

• JavaScript's window.getComputedStyle() method

• Eric Meyer and Aaron Gustafson's matrix resolution tool

How to use window.getComputedStyle() to convert transform functions to matrix()

Suppose you want to convert the following transform functions to matrix:

img {
  transform-origin: 0 0;
  transform: scale(2) translateX(100px);
  width: 80%;
}

You will add an id attribute to the image element:

<img
  src="https://cdn.pixabay.com/photo/2022/09/26/23/26/african-american-7481724_960_720.jpg"
  alt=""
  id="image"
/>

Then, in JavaScript, you will:

1. Use the id attribute to get the image element.

2. Use the window.getComputedStyle() method to get the image's transform property's value.

Here's the code:

// Get the image element by its id name:
const image = document.getElementById("image");

// Get the image element's transform property's value:
const matrix = window.getComputedStyle(image).getPropertyValue("transform");

// Log the matrix variable's value to the console:
console.log(matrix);

Try it on StackBlitz

Browsers, by default, convert a CSS transform property's value to its matrix equivalent. So, the snippet above returned the image's computed value.

Let's now discuss the second conversion tool.

How to use the matrix resolutions tool to convert transform functions to matrix()

Suppose you want to convert the following transform functions to a matrix():

img {
  transform-origin: 0 0;
  transform: scale(2) translateX(100px);
  width: 80%;
}

You will do the following:

1. Go to The Matrix Resolutions website: https://meyerweb.com/eric/tools/matrix/.

2. Paste your transform functions (scale(2) translateX(100px)) into the first text field.

3. Click "The Red Pill" button to generate the transform functions' matrix equivalence.

Click the red pill button to convert CSS transform functions to a matrix() function

Tip: Use matrix3d() to create a 3D transformation matrix.

Important Stuff to Know about Transforming Elements in CSS

Here are three essential facts to remember when you transform elements in CSS.

1. Transform creates a stacking context

Suppose you set the transform property to any value other than none. In that case, the browser will create a stacking context. And the transformed element will serve as a containing block to any absolute or fixed positioned elements it contains.

2. Scaling and zooming animations cause accessibility issues

Whenever you include scaling or zooming animations in your app, provide users an option to turn off animations. This option is necessary because scaling and zooming animations cause accessibility issues.

3. Not all elements are transformable

You cannot transform the following box models:

• Non-replaced inline elements

• table-column boxes

• table-column-group

Wrapping Up

In this article, we discussed all the CSS transform tools you need to translate, rotate, skew, scale, or add perspective effects to HTML elements.

I hope you've found this article helpful.

Thanks for reading!

If you like this tutorial, you will enjoy my CSS Flexbox book. It is a handy quick reference guide that uses images and live examples to explain Flexbox.

Is Coursera Free? At Class Central, we get that question so often that I wrote a guide to answer it. Nowadays, most Coursera courses are free to audit. But if you want to complete graded assignments or get a course certificate, you have to pay.

Coursera’s entire catalog used to be entirely free. But in late 2015, they introduced a paywall. Over the following years, this monetization approach expanded to other online course providers, including edX.

But I always knew that there were a few courses on Coursera that remained completely free. When you enroll in these courses, they offer a “Full Course, No Certificate” option, as you can see below.

Free Coursera course: “Full Course, No Certificate” option (with graded assignments)

By contrast, other courses tend to offer a more restrictive free-to-audit option, which doesn’t include graded assignments. This option is also quite easy to miss. You have to look for the “audit” link at the very bottom of the window that pops up when you enroll.

Free Coursera course: Free audit option (without graded assignments)

Which Courses are Entirely Free?

Out of Coursera’s 7000 or so active courses, I expected to find about 50 fully-free courses. Imagine my surprise (shock, really) when I found that close to 1700 Coursera courses are still completely free

The biggest challenge is finding which Coursera courses are entirely free. The Coursera website doesn’t provide any “free course” filter. To know if a course is fully free, you have to log into Coursera, enroll in the course, and look for the “Full Course, No Certificate” option.

But there are thousands of courses on Coursera, so short of checking them one by one, it’s tricky to manually put together a list of their fully-free courses. So instead, we went with an automated approach.

Originally, I used the browser automation tool Selenium to check the courses for me. My colleague Bob set me up with an AWS instance and Cloud9 IDE, where I could do this automation. Below is a screenshot of the setup. (Yes, I prefer PHP.)

Automating the search for free Coursera courses with Selenium

Out of Coursera’s 7000 or so active courses, I expected to find about 50 fully-free courses. Imagine my surprise (shock, really) when I found that close to 1700 Coursera courses are still completely free (that is, including the graded assignments, minus only the free certificate).

Free Coursera Courses

Below, I've compiled the 1700 entirely free courses offered on Coursera. It includes a number of highly-rated courses, including a few from Class Central’s Best Online Courses of All Time, such as Barbara Oakley's Learning How to Learn.

Here are some stats about the course list:

• Combined, these courses have 76M enrollments.
• Combined, they have received 41K reviews on Class Central.
• There are 9 courses with over 1M enrollments.
• The most popular one, Yale's Science of Well-Being, has 4.3M enrollments.
• 58% of the courses are in English and 14% are in Spanish.

To help you navigate the course list, I’ve broken it down by subject. Click on one to jump to the corresponding courses.

• Personal Development (78)
• Health & Medicine (196)
• Science (194)
• Humanities (197)
• Computer Science (103)
• Programming (81)
• Mathematics (44)
• Engineering (88)
• Social Sciences (212)
• Business (255)
• Art & Design (67)
• Data Science (45)
• Education & Teaching (85)
• Information Security (InfoSec) (16)

More Free Courses

If you don’t find what you need in the course list below, you may want to have a look at my other freeCodeCamp articles:

• 1000+ Free Developer Certifications
• Here are 850+ Ivy League Courses You Can Take Right Now for Free
• 860+ Free Online Programming & Computer Science Courses

Personal Development (78)

• Learning How to Learn: Powerful mental tools to help you master tough subjects from Deep Teaching Solutions ★★★★★(23000)
• Mindshift: Break Through Obstacles to Learning and Discover Your Hidden Potential from McMaster University ★★★★★(7364)
• Learning How To Learn for Youth from Arizona State University ★★★★★(948)
• Apprendre comment apprendre (ACA) : Des outils mentaux puissants qui vous aideront à maîtriser des sujets difficiles from Deep Teaching Solutions ★★★★★(377)
• Successful Negotiation: Essential Strategies and Skills from University of Michigan ★★★★★(217)
• Aprender from Universidad Nacional Autónoma de México ★★★★★(106)
• Know Thyself - The Value and Limits of Self-Knowledge: The Examined Life from University of Edinburgh ★★★★★(30)
• How to Get Skilled: Introduction to Individual Skills Management (Project-Centered Course) from State University of New York ★★★★☆(8)
• Mindware: Critical Thinking for the Information Age from University of Michigan ★★★★★(6)
• The Arts and Science of Relationships: Understanding Human Needs from University of Toronto ★★★★☆(5)
• Aprendendo a aprender: ferramentas mentais poderosas para ajudá-lo a dominar assuntos difíceis (em Português) [Learning How to Learn] from Deep Teaching Solutions ★★★★☆(5)
• Competencias digitales. Herramientas de ofimática (Microsoft Word, Excel, Power Point) from Universitat Autònoma de Barcelona (Autonomous University of Barcelona) ★★★★★(3)
• Digital Footprint from University of Edinburgh ★★★★★(2)
• Transmedia Storytelling: Narrative worlds, emerging technologies, and global audiences from University of New South Wales ★★★☆☆(2)
• Operating System Foundations from Infosec ★★☆☆☆(2)
• تعّلم كيف تتعلم: أدوات ذهنية قوية لمساعدتك على إتقان موضوعات صعبة from Deep Teaching Solutions ★★★★★(2)
• 学会如何学习：帮助你掌握复杂学科的强大智力工具（Learning How to Learn） from Deep Teaching Solutions ★★★★★(2)
• Managing Emotions in Times of Uncertainty & Stress from Yale University ★★★★★(1)
• The Nature of Genius from Yale University ★★★★★(1)
• Lesson | Small Talk & Conversational Vocabulary from Georgia Institute of Technology ★★★★★(1)
• Mastering Remote Work and Online Study in U.S. in the post-COVID Era from State University of New York ★★★★★(1)
• Job Success: Get Hired or Promoted in 3 Steps from State University of New York ★☆☆☆☆(1)
• Understanding and Managing the Stresses of Police Work from University of Toronto ★★★☆☆(1)
• "Making" Progress Teach-Out from Emory University ★★★★★(1)
• Trabajar para ser feliz from Universidad Austral ★★★☆☆(1)
• Research for Impact from University of Cape Town ★★★★★(1)
• Translation in Practice from Nanjing University ★★★☆☆(1)
• English for Media Literacy from University of Pennsylvania
• علم النجاح: ما يجب أن تعرفه من الباحثين from University of Michigan
• Brilliant, Passionate You from University of Michigan
• Réussir une négociation : Stratégies et compétences clés from University of Michigan
• Erfolgreich Verhandeln: Strategien und Fähigkeiten from University of Michigan
• التأثير على الأشخاص from University of Michigan
• Como influenciar pessoas from University of Michigan
• Negociación exitosa: Estrategias y habilidades esenciales (en español) from University of Michigan
• Negociações de sucesso: estratégias e habilidades essenciais (em Português) from University of Michigan
• International Travel Preparation, Safety, & Wellness from Johns Hopkins University
• Introduction to Learning Transfer and Life Long Learning (3L) from University of California, Irvine
• Працюйте розумніше, а не більше: управління часом для особистої та професійної продуктивності from University of California, Irvine
• مقدمة عن التفاوض: دليل استراتيجي لتصبح مُفاوضًا ذا مبادئ ومُقنعًا from Yale University
• Введение: как стать убедительным переговорщиком from Yale University
• Lesson | Express Yourself: Pronunciation from Georgia Institute of Technology
• Lesson | Organize Your Pitch from Georgia Institute of Technology
• Lesson | Get Ready for the Interview from Georgia Institute of Technology
• Lesson | Understand and Be Understood on the Phone from Georgia Institute of Technology
• Lesson | Telephone Language from Georgia Institute of Technology
• Lesson | Video Conferencing: Face to Face but Online from Georgia Institute of Technology
• الخطابة الإقناعية: تحفيز الجماهير بالحجج المُقنِعة واللغة المؤثِّرة from University of Washington
• Trabajo en equipo y creatividad from University of Illinois at Urbana-Champaign
• اللغة والمفاهيم التنظيمية from University of Illinois at Urbana-Champaign
• Travail d'équipe et créativité from University of Illinois at Urbana-Champaign
• العمل الجماعي والإبداع from University of Illinois at Urbana-Champaign
• Empowering Yourself in a Post-Truth World from State University of New York
• Power Onboarding from Northwestern University
• Exercer son esprit critique à l’ère informationnelle from University of Geneva
• Designing the Future of Work from University of New South Wales
• Conflict Transformation from Emory University
• Les Fondamentaux de la Négociation from ESSEC Business School
• 職場素養 (Professionalism) from National Taiwan University
• La solución del conflicto ético from Universidad Nacional Autónoma de México
• Competencias digitales de ofimática: Microsoft Access from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Competencias digitales. Conceptos y herramientas básicas from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• La formación del pensamiento crítico en la era digital from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Cómo hablar bien en público from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Pensamiento Sistémico en el Siglo XXI from Universidad de los Andes
• Competencias para la empleabilidad from Universidad Austral
• Negociación 4.0 from Universidad Austral
• Повествование и влияние: Эффектная коммуникация from Macquarie University
• Storytelling et influence : Communiquer pour convaincre from Macquarie University
• Data Analyst Career Guide and Interview Preparation from IBM
• Science Communication: Communicating Trustworthy Information in the Digital World from Erasmus University Rotterdam
• Mi Primer Empleo (MPE) from Universidad de Chile
• Etkili Konuşma (Effective Speaking) from Koç University
• Professionalism in an era of change from Utrecht University
• Como engajar equipes e criar cultura em ambientes virtuais from Fundação Instituto de Administração
• التفكير الناقد | Critical Thinking from Alfaisal University
• A tanulás tanulása: Hatékony mentális eszközök, melyek segítenek megbirkózni a nehéz tantárgyakkal (Learning How to Learn) from Deep Teaching Solutions
• Land Your Next Role from Aptly

Health & Medicine (196)

• Understanding Clinical Research: Behind the Statistics from University of Cape Town ★★★★★(735)
• Understanding Medical Research: Your Facebook Friend is Wrong from Yale University ★★★★★(364)
• Health Concepts in Chinese Medicine from The Hong Kong University of Science and Technology ★★★★★(86)
• Vital Signs: Understanding What the Body Is Telling Us from University of Pennsylvania ★★★★★(51)
• Epidemiology: The Basic Science of Public Health from The University of North Carolina at Chapel Hill ★★★★☆(29)
• Severe to Profound Intellectual Disability: Circles of Care and Education from University of Cape Town ★★★★★(24)
• Positive Psychiatry and Mental Health from The University of Sydney ★★★★★(21)
• COVID-19 Contact Tracing from Johns Hopkins University ★★★★★(16)
• Sit Less, Get Active from University of Edinburgh ★★★★☆(16)
• Organ Donation: From Death to Life from University of Cape Town ★★★★★(14)
• Mind Control: Managing Your Mental Health During COVID-19 from University of Toronto ★★★★★(12)
• Epidemics - the Dynamics of Infectious Diseases from Pennsylvania State University ★★★★☆(11)
• COVID-19: What You Need to Know from Osmosis University ★★★★★(11)
• Clinical Terminology for International and U.S. Students from University of Pittsburgh ★★★★★(10)
• COVID-19 Training for Healthcare Workers from Stanford University ★★★★★(9)
• Introduction to Dental Medicine from University of Pennsylvania ★★★★☆(9)
• Infection Prevention in Nursing Homes from The University of North Carolina at Chapel Hill ★★★★★(9)
• Bacteria and Chronic Infections from University of Copenhagen ★★★★★(8)
• Managing Your Health: The Role of Physical Therapy and Exercise from University of Toronto ★★★★☆(7)
• Diabetes - a Global Challenge from University of Copenhagen ★★★★☆(7)
• Chemicals and Health from Johns Hopkins University ★★★★☆(6)
• Preventing Chronic Pain: A Human Systems Approach from University of Minnesota ★★★★★(6)
• Science Matters: Let's Talk About COVID-19 from Imperial College London ★★★★★(6)
• Health Across the Gender Spectrum from Stanford University ★★★★★(5)
• Fighting COVID-19 with Epidemiology: A Johns Hopkins Teach-Out from Johns Hopkins University ★★★★★(5)
• Systems Thinking In Public Health from Johns Hopkins University ★★★★★(5)
• EDIVET: Do you have what it takes to be a veterinarian? from University of Edinburgh ★★★★☆(5)
• The Social Context of Mental Health and Illness from University of Toronto ★★★☆☆(5)
• Disease Screening in Public Health from University of Geneva ★★★★☆(5)
• Data Management for Clinical Research from Vanderbilt University ★★★★☆(5)
• Anatomy of the Abdomen and Pelvis; a journey from basis to clinic. from Leiden University ★★★★★(5)
• The New Nordic Diet - from Gastronomy to Health from University of Copenhagen ★★★★☆(5)
• Stanford's Short Course on Breastfeeding from Stanford University ★★★★★(4)
• Teaching and Assessing Clinical Skills from University of Michigan ★★★★★(4)
• Measuring and Maximizing Impact of COVID-19 Contact Tracing from Johns Hopkins University ★★★★☆(4)
• ADHD: Everyday Strategies for Elementary Students from State University of New York ★★★★★(4)
• Ebola: Essential Knowledge for Health Professionals from University of Amsterdam ★★★★★(4)
• Antibiotic Stewardship from Stanford University ★★★★★(3)
• Michigan Sport-Related Concussion Training Certification from University of Michigan ★★★★☆(3)
• AIDS: Fear and Hope from University of Michigan ★★★★★(3)
• Strategies for Senior Housing Communities during COVID-19 from Johns Hopkins University ★★★★★(3)
• Anatomy of the Chest, Neck, Abdomen, and Pelvis from Yale University ★★★★★(3)
• Bats, Ducks, and Pandemics: An Introduction to One Health Policy from Princeton University ★★★★★(3)
• Tropical Parasitology: Protozoans, Worms, Vectors and Human Diseases from Duke University ★★★★★(3)
• Foundations for Assisting in Home Care from State University of New York ★★★★★(3)
• Chicken Behaviour and Welfare from University of Edinburgh ★★★☆☆(3)
• Global Health: An Interdisciplinary Overview from University of Geneva ★★★★☆(3)
• Ebola Virus Disease: An Evolving Epidemic from Emory University ★★★★☆(3)
• Clinical Kidney, Pancreas and Islet Transplantation from Leiden University ★★★★☆(3)
• Everyday Chinese Medicine from The Chinese University of Hong Kong ★★★★☆(3)
• Implant Dentistry from The University of Hong Kong ★★★★★(3)
• An Introduction to Global Health from University of Copenhagen ★★★★☆(3)
• Introduction to Cataract Surgery from University of Michigan ★★★★☆(2)
• COVID Vaccine Ambassador Training: How to Talk to Parents from Johns Hopkins University ★★★★★(2)
• Addiction Treatment: Clinical Skills for Healthcare Providers from Yale University ★★★☆☆(2)
• Health Behavior Change: From Evidence to Action from Yale University ★★★★★(2)
• Essentials of Global Health from Yale University ★★★★☆(2)
• Career 911: Your Future Job in Medicine and Healthcare from Northwestern University ★★★★★(2)
• Childbirth: A Global Perspective from Emory University ★★★★★(2)
• Actualización en el manejo del paciente con diabetes mellitus tipo 2 from Universidad Nacional Autónoma de México ★★★★★(2)
• 杏林探宝——认知中药 from Shanghai Jiao Tong University ★★★★★(2)
• Everyday Chinese Medicine 2 from The Chinese University of Hong Kong ★★★★★(2)
• Diabetes – the Essential Facts from University of Copenhagen ★★★★★(2)
• Easing the burden of obesity, diabetes and cardiovascular disease from The University of Sydney ★★★★★(2)
• MRI Fundamentals from Korea Advanced Institute of Science and Technology ★★★☆☆(2)
• Health Care IT: Challenges and Opportunities from Icahn School of Medicine at Mount Sinai ★★★☆☆(2)
• Clinical Epidemiology from Utrecht University ★★★★★(2)
• Capacitación sobre COVID-19 para trabajadores de salud from Stanford University ★★★★★(1)
• Stories of Infection from Stanford University ★★★★★(1)
• The Oral Cavity: Portal to Health and Disease from University of Pennsylvania ★★★★★(1)
• Thoracic Oncology from University of Michigan ★★★★☆(1)
• Public Health in Humanitarian Crises 2 from Johns Hopkins University ★★★★★(1)
• Strategies for Assisted Living Communities during COVID-19 from Johns Hopkins University ★★★★★(1)
• Overcoming Dyslexia from Yale University ★★★★★(1)
• Global Health Diplomacy from State University of New York ★★★★★(1)
• Precision Medicine from University of Geneva ★★★★★(1)
• Case Studies in Personalized Medicine from Vanderbilt University ★★★★★(1)
• Introduction to Household Water Treatment and Safe Storage from École Polytechnique Fédérale de Lausanne ★★★★★(1)
• 食品安全與毒理 (Food Safety & Toxicology) from National Taiwan University ★★☆☆☆(1)
• 食品安全與風險分析（Food Safety＆Risk Analysis） from National Taiwan University ★★★★★(1)
• Nutrición y obesidad: control de sobrepeso from Universidad Nacional Autónoma de México ★★★★★(1)
• Actúa ante el dolor crónico from Universitat Autònoma de Barcelona (Autonomous University of Barcelona) ★★★★★(1)
• 口腔种植学 （Implant Dentistry） from The University of Hong Kong ★★★★★(1)
• Urbanisation and Health - Promoting Sustainable Solutions from University of Copenhagen ★★★★★(1)
• Innovating Solutions for Aging Populations from University of Copenhagen ★★★★★(1)
• Prehospital care of acute stroke and patient selection for endovascular treatment using the RACE scale from University of Barcelona ★★★★★(1)
• Alimentación y dietética para una vuelta al mundo a vela from University of Barcelona ★★★★★(1)
• eHealth: More than just an electronic record from The University of Sydney ★★★★☆(1)
• Leading Healthcare Quality and Safety from George Washington University ★★★★★(1)
• Cuidados y procedimientos generales en la atención del recién nacido from Universidad de Chile ★☆☆☆☆(1)
• Medical Applications of Particle Accelerators (NPAP MOOC) from Lund University ★★★☆☆(1)
• İnfertilite Hemşireliği (Infertility Nursing) from Koç University ★★★★★(1)
• Lecture Series for Preventing and Controlling COVID-19 from Xi'an Jiaotong University ★★★★★(1)
• स्वास्थ्य कर्मचारियों के लिए कोविड-19 ट्रेनिंग from Stanford University
• Health After Cancer: Cancer Survivorship for Primary Care from Stanford University
• Formação sobre a COVID-19 para Profissionais de Saúde from Stanford University
• Formation COVID-19 pour personnels de santé from Stanford University
• Teaching LGBTQ+ Health from Stanford University
• Feeding the World from University of Pennsylvania
• Hearing Loss in Children from University of Michigan
• Service Transformed: Lessons in U.S. Veteran Centered Care from University of Michigan
• Rastreamento de contato da COVID-19 from Johns Hopkins University
• Rastreo de los contactos de la COVID-19 from Johns Hopkins University
• Let's talk about it: A Health and Immigration Teach Out from Johns Hopkins University
• Embajador de la Vacuna COVID: Cómo Hablar con los Padres from Johns Hopkins University
• Global Quality Maternal and Newborn Care from Yale University
• Planetary Health for Nurses from Yale University
• The Science of Well-Being for Teens from Yale University
• Hacking COVID-19 — Course 3: Unraveling COVID-19's Origins from University of California, San Diego
• Mental Health and Resilience for Healthcare Workers from University of Toronto
• 2022 Cancer Prevention Web-Based Activity from University of Virginia
• Global Adolescent Health from University of Melbourne
• 流行病学基础（上） from Peking University
• 社会调查与研究方法 （上）Methodologies in Social Research (Part I) from Peking University
• 社会调查与研究方法 （下）Methodologies in Social Research (Part 2) from Peking University
• 更年期综合管理 from Peking University
• 大学生瑜伽 from Peking University
• Community Awareness Course: Sexuality and Disability from University of Minnesota
• Supervision du raisonnement clinique from University of Geneva
• Global Health Security, Solidarity and Sustainability through the International Health Regulations from University of Geneva
• Troubles du spectre de l'autisme : diagnostic from University of Geneva
• COVID-19 : à la recherche des contacts from University of Geneva
• Troubles du spectre de l'autisme : biologie et neurosciences from University of Geneva
• Troubles du spectre de l'autisme : interventions from University of Geneva
• Ebola : Vaincre ensemble ! from University of Geneva
• In the footsteps of Zika… approaching the unknown from University of Geneva
• Violence Against Healthcare from University of Geneva
• Faster Together, Enhancing the Recruitment of Minorities in Clinical Trials from Vanderbilt University
• Health, Society, and Wellness in COVID-19 Times from University of Colorado Boulder
• #talkmentalillness from University of Colorado Boulder
• Foundations for Global Health Responders from University of Colorado System
• Introduction to Public Health Engineering in Humanitarian Contexts from École Polytechnique Fédérale de Lausanne
• Pioneers of Medicine and Medical Breakthroughs in Taiwan from National Taiwan University
• 一堂課讓你認識肺癌（Basic Concepts of Lung Cancer: Diagnosis and Treatment） from National Taiwan University
• Diálisis peritoneal from Universidad Nacional Autónoma de México
• Farmacología para odontólogos from Universidad Nacional Autónoma de México
• Farmacovigilancia ocular from Universidad Nacional Autónoma de México
• Cuidado de heridas en el ámbito hospitalario from Universidad Nacional Autónoma de México
• Manejo moderno de la caries dental from Universidad Nacional Autónoma de México
• Cáncer de próstata from Universidad Nacional Autónoma de México
• El rol docente del médico residente from Universidad Nacional Autónoma de México
• Evaluación del Aprendizaje en Escenarios Clínicos from Tecnológico de Monterrey
• Population Health: Fundamentals of Population Health Management from Leiden University
• Population Health: Governance from Leiden University
• Population Health: Health & Health Behaviour from Leiden University
• Population Health: Study Design from Leiden University
• Population Health: Responsible Data Analysis from Leiden University
• Population Health: Alternative Payment Models from Leiden University
• Population Health: Panel Management Next Level from Leiden University
• Anticoncepción hormonal al alcance de todos from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Primeros Auxilios Psicológicos (PAP). Edición especial COVID-19 from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Curanderismo:Influencias Globales de la Medicina Tradicional from University of New Mexico
• Tools and Practices for Addressing Pandemic Challenges from Politecnico di Milano
• Trustworthy AI for Healthcare Management from Politecnico di Milano
• Reaching Vulnerable Groups through Pandemic Policy from Politecnico di Milano
• Diagnóstico y tratamiento del dolor neuropático en atención primaria from Universidad de los Andes
• Anticoagulación from Universidad de los Andes
• COVID-19 - A clinical update from University of Florida
• Biomedical Visualisation from University of Glasgow
• The Social Dimensions of Antimicrobial Resistance from University of Amsterdam
• 循序渐进练瑜伽 from Shanghai Jiao Tong University
• Improving Immunity Based on Traditional Eastern Exercises from Shanghai Jiao Tong University
• 常见慢性病的健康管理 from Shanghai Jiao Tong University
• Materials in Oral Health from The University of Hong Kong
• Epidemics from The University of Hong Kong
• Global Health Policy from University of Tokyo
• Business Models for Innovative Care for Older People from University of Copenhagen
• Non-Communicable Diseases in Humanitarian Settings from University of Copenhagen
• Understanding Patient Perspectives on Medications from University of Copenhagen
• Non-Alcoholic Fatty Liver Disease – The Essentials from University of Copenhagen
• Patient Perspectives on Medications: Qualitative Interviews from University of Copenhagen
• 抗菌药与超级细菌 －－天使与魔鬼的博弈 from Fudan University
• 全球卫生导论 from Fudan University
• Atención prehospitalaria del ictus agudo y selección de pacientes para tratamiento endovascular con la escala RACE from University of Barcelona
• ÉCHELLE RACE: Dépistage préhospitalier de l’AVC impliquant une occlusion de gros vaisseaux cérébraux from University of Barcelona
• Manejo del enfermo semicrítico y crítico por COVID-19 from University of Barcelona
• Global Health and Humanitarianism from University of Manchester
• Doing Clinical Research: Biostatistics with the Wolfram Language from University of Cape Town
• The Basics of Trauma Surgery from Technische Universität München (Technical University of Munich)
• Grundlagen der Unfallchirurgie from Technische Universität München (Technical University of Munich)
• Essentials in Clinical Simulations Across the Health Professions from George Washington University
• idea 2 IMPACT: An Introduction to Translating Assistive Health Technologies and Other Products from University of Pittsburgh
• Meditation: A way to achieve your goals in your life from Korea Advanced Institute of Science and Technology
• Manejo de Crisis Epilépticas y Promoción de la Inclusión Social from Pontificia Universidad Católica de Chile
• Recuperación Optimizada en Cirugía Colorrectal from Pontificia Universidad Católica de Chile
• Sistemas y modelos de organización de la atención en salud from Pontificia Universidad Católica de Chile
• Atención Primaria en Salud: El desafío de las Enfermedades no Transmisibles from Pontificia Universidad Católica de Chile
• Regulated Competition in Healthcare Systems: Theory & Practice from Erasmus University Rotterdam
• Circadian clocks: how rhythms structure life from Ludwig-Maximilians-Universität München
• 营养与健康 from Nanjing University
• Virology and Epidemiology in the Time of COVID-19 from American Museum of Natural History
• Revisão Sistemática e Meta-análise from Universidade Estadual de Campinas
• Cosmetologia em prol do envelhecimento com saúde e beleza from Universidade Estadual de Campinas
• Cosmetologia: Prática de Manipulação de Cosméticos from Universidade Estadual de Campinas
• Saúde Baseada em Evidências from Universidade Estadual de Campinas
• Engaging in Strategic Management in Healthcare Organizations from Northeastern University

Science (194)

• Mountains 101 from University of Alberta ★★★★★(597)
• Dino 101: Dinosaur Paleobiology from University of Alberta ★★★★★(139)
• Medical Neuroscience from Duke University ★★★★★(51)
• The Science of the Solar System from California Institute of Technology ★★★★★(51)
• Paleontology: Theropod Dinosaurs and the Origin of Birds from University of Alberta ★★★★★(42)
• Astrobiology and the Search for Extraterrestrial Life from University of Edinburgh ★★★★★(33)
• Introduction to Genetics and Evolution from Duke University ★★★★★(31)
• Climate Change Mitigation in Developing Countries from University of Cape Town ★★★★★(30)
• Introductory Human Physiology from Duke University ★★★★☆(29)
• Paleontology: Ancient Marine Reptiles from University of Alberta ★★★★★(29)
• How Things Work: An Introduction to Physics from University of Virginia ★★★★★(28)
• Understanding the Brain: The Neurobiology of Everyday Life from The University of Chicago ★★★★★(28)
• Paleontology: Early Vertebrate Evolution from University of Alberta ★★★★★(27)
• Astronomy: Exploring Time and Space from University of Arizona ★★★★★(23)
• Understanding Einstein: The Special Theory of Relativity from Stanford University ★★★★★(19)
• Dog Emotion and Cognition from Duke University ★★★★☆(19)
• From the Big Bang to Dark Energy from University of Tokyo ★★★★☆(17)
• Introduction to Forensic Science from Nanyang Technological University ★★★★☆(17)
• Animal Behaviour and Welfare from University of Edinburgh ★★★★☆(13)
• Disaster Preparedness from University of Pittsburgh ★★★★☆(13)
• Introduction to Physical Chemistry from University of Manchester ★★★★☆(12)
• An Introduction to Consumer Neuroscience & Neuromarketing from Copenhagen Business School ★★★☆☆(12)
• AstroTech: The Science and Technology behind Astronomical Discovery from University of Edinburgh ★★★★☆(11)
• The Truth About Cats and Dogs from University of Edinburgh ★★★★★(11)
• Introduction to Human Behavioral Genetics from University of Minnesota ★★★★★(11)
• Understanding Plants - Part I: What a Plant Knows from Tel Aviv University ★★★★★(11)
• Exploring Quantum Physics from University of Maryland, College Park ★★★★☆(10)
• Gut Check: Exploring Your Microbiome from University of Colorado Boulder ★★★★☆(10)
• The Bilingual Brain from University of Houston System ★★★★☆(10)
• Synapses, Neurons and Brains from Hebrew University of Jerusalem ★★★★☆(10)
• Introduction to Chemistry: Reactions and Ratios from Duke University ★★★★★(9)
• Computational Neuroscience from University of Washington ★★★★☆(9)
• Astro 101: Black Holes from University of Alberta ★★★★★(9)
• Global Warming I: The Science and Modeling of Climate Change from The University of Chicago ★★★★☆(9)
• Water Supply and Sanitation Policy in Developing Countries Part 1: Understanding Complex Problems from University of Manchester ★★★★★(8)
• Ecology: Ecosystem Dynamics and Conservation from American Museum of Natural History ★★★★★(8)
• Advanced Chemistry from University of Kentucky ★★★★★(8)
• Origins - Formation of the Universe, Solar System, Earth and Life from University of Copenhagen ★★★★★(7)
• Greening the Economy: Lessons from Scandinavia from Lund University ★★★★☆(7)
• Visual Perception and the Brain from Duke University ★★★★☆(6)
• Archaeoastronomy from Politecnico di Milano ★★★★★(6)
• Analyzing the Universe from Rutgers University ★★★★☆(6)
• Introduction to Molecular Spectroscopy from University of Manchester ★★★★☆(6)
• Agriculture, Economics and Nature from University of Western Australia ★★★★★(6)
• Introduction to Thermodynamics: Transferring Energy from Here to There from University of Michigan ★★★★☆(5)
• Sleep: Neurobiology, Medicine, and Society from University of Michigan ★★★★★(5)
• Materials Science: 10 Things Every Engineer Should Know from University of California, Davis ★★★★☆(5)
• General Chemistry: Concept Development and Application from Rice University ★★★★☆(5)
• A Circular Economy of Metals: Towards a Sustainable Societal Metabolism from Leiden University ★★★★★(5)
• Our Earth: Its Climate, History, and Processes from University of Manchester ★★★★★(5)
• The Finite Element Method for Problems in Physics from University of Michigan ★★★★☆(4)
• Introduction to Chemistry: Structures and Solutions from Duke University ★★★★☆(4)
• Music as Biology: What We Like to Hear and Why from Duke University ★★★★☆(4)
• Introduzione alla fisica sperimentale: meccanica, termodinamica from Politecnico di Milano ★★★★★(4)
• Understanding Plants - Part II: Fundamentals of Plant Biology from Tel Aviv University ★★★★☆(4)
• Imagining Other Earths from Princeton University ★★★★☆(3)
• Pensamiento científico from Universidad Nacional Autónoma de México ★★★☆☆(3)
• The Horse Course: Introduction to Basic Care and Management from University of Florida ★★★★★(3)
• The Science of Stem Cells from American Museum of Natural History ★★★★★(3)
• Statistical Mechanics: Algorithms and Computations from École normale supérieure ★★★★★(3)
• Emergent Phenomena in Science and Everyday Life from University of California, Irvine ★★★★☆(2)
• Material Behavior from Georgia Institute of Technology ★★★☆☆(2)
• Sustainable Food Production Through Livestock Health Management from University of Illinois at Urbana-Champaign ★★★★★(2)
• Chemical Biology from University of Geneva ★★★★★(2)
• The Diversity of Exoplanets from University of Geneva ★★★★★(2)
• Water Resources Management and Policy from University of Geneva ★★★☆☆(2)
• Thermodynamique : fondements from École Polytechnique Fédérale de Lausanne ★★★★★(2)
• Air Pollution – a Global Threat to our Health from University of Copenhagen ★★★★★(2)
• Data-driven Astronomy from The University of Sydney ★★★★★(2)
• Discover Best Practice Farming for a Sustainable 2050 from University of Western Australia ★★★★☆(2)
• Experimentation for Improvement from McMaster University ★★★★★(2)
• Evolution: A Course for Educators from American Museum of Natural History ★★★★★(2)
• The Dynamic Earth: A Course for Educators from American Museum of Natural History ★★★★☆(2)
• Chimpanzee Behavior and Conservation from Duke University ★★★★★(1)
• Global Postharvest Loss Prevention: Fundamentals, Technologies, and Actors from University of Illinois at Urbana-Champaign ★★★★★(1)
• Introduction to Reproduction from Northwestern University ★★★★★(1)
• Advanced Neurobiology I from Peking University ★★★☆☆(1)
• Advanced Neurobiology II from Peking University ★☆☆☆☆(1)
• Global Health at the Human-Animal-Ecosystem Interface from University of Geneva ★★★★★(1)
• The Sun and the Total Eclipse of August 2017 from University of Colorado Boulder ★★★☆☆(1)
• Mechanics: Motion, Forces, Energy and Gravity, from Particles to Planets from University of New South Wales ★★★★☆(1)
• Mécanique de Newton from École Polytechnique Fédérale de Lausanne ★★★★★(1)
• Mécanique du point matériel from École Polytechnique Fédérale de Lausanne ★★★★☆(1)
• Introduction to Faecal Sludge Management from École Polytechnique Fédérale de Lausanne ★★★☆☆(1)
• Las estaciones del año y el clima from Universidad Nacional Autónoma de México ★★★★★(1)
• ¡¿Cómo?! ¿Química en mi casa? from Universidad Nacional Autónoma de México ★★★★★(1)
• Evolution Today from Leiden University ★★★★★(1)
• Fundamentals of Fluid-Solid Interactions from École Polytechnique ★★★★★(1)
• Mécanique quantique from École Polytechnique ★★★★★(1)
• Introduzione alla fisica sperimentale: elettromagnetismo, ottica, fisica moderna from Politecnico di Milano ★★★★★(1)
• Introduction to the Arctic: Climate from University of Alberta ★★★★★(1)
• Science & Religion 101 from University of Alberta ★★★★★(1)
• Sustainable Agricultural Land Management from University of Florida ★★★★★(1)
• The Evolving Universe from California Institute of Technology ★★★★☆(1)
• Transformation of the Global Food System from University of Copenhagen ★★★★★(1)
• Origens da Vida no Contexto Cósmico from Universidade de São Paulo ★☆☆☆☆(1)
• Confronting The Big Questions: Highlights of Modern Astronomy from University of Rochester ★★★★☆(1)
• Oceanography: a key to better understand our world from University of Barcelona ★☆☆☆☆(1)
• Estrategias para ganar. Meteorología en la vuelta al mundo from University of Barcelona ★☆☆☆☆(1)
• Industrial Biotechnology from University of Manchester ★★★★★(1)
• From Climate Science to Action from Online Learning Campus - World Bank Group ★★★☆☆(1)
• A Brief History of Human Spaceflight from University of Houston System ★★★★☆(1)
• Climate Change and Human Rights from New York University (NYU) ★★★★★(1)
• Patenting in Biotechnology from Technical University of Denmark (DTU) ★★★★★(1)
• Global Warming II: Create Your Own Models in Python from The University of Chicago ★★☆☆☆(1)
• Introduction to Acoustics (Part 2) from Korea Advanced Institute of Science and Technology ★★☆☆☆(1)
• Astrobiology: Exploring Other Worlds from University of Arizona ★★★★★(1)
• Biosphere 2 Science for the Future of Our Planet from University of Arizona ★★★★★(1)
• 结构生物化学（Structural Biochemistry) from Nanjing University ★★★★★(1)
• Wood Science: Beyond Building from West Virginia University ★★★★★(1)
• Exploring Light: Hands-on Activities and Strategies for Teachers from Exploratorium ★★★★☆(1)
• Black Agricultural Solutions to Food Apartheid: A Teach-Out from Johns Hopkins University
• Tropical Forest Landscapes 101: Conservation & Restoration from Yale University
• 宇宙之旅：对话 (Journey of the Universe: Weaving Knowledge and Action) from Yale University
• Thomas Berry的世界观：地球社区的繁荣 (The Worldview of Thomas Berry: The Flourishing of the Earth Community) from Yale University
• 宇宙之旅：展现生命 (Journey of the Universe: The Unfolding of Life) from Yale University
• Algae Biotechnology from University of California, San Diego
• Introduction to Seaweeds from University of California, San Diego
• Introduction to Algae from University of California, San Diego
• Space Medicine from Duke University
• 人体生理学导论（中文版） from Duke University
• Materials Data Sciences and Informatics from Georgia Institute of Technology
• 电磁学上——恒定电场 from Peking University
• 生物演化 from Peking University
• 大学化学 from Peking University
• 电磁学下——恒磁场与时变电磁场 from Peking University
• 遗传学实验 from Peking University
• 生物学概念与途径 from Peking University
• Understanding Modern Physics I: Relativity and Cosmology from The Hong Kong University of Science and Technology
• Understanding Modern Physics III: Simplicity and Complexity from The Hong Kong University of Science and Technology
• Physique des particules - une introduction from University of Geneva
• Classical papers in molecular genetics from University of Geneva
• Climate Change and Water in Mountains: A Global Concern from University of Geneva
• Genetic Epidemiology Foundations from University of Colorado System
• Transmission electron microscopy for materials science from École Polytechnique Fédérale de Lausanne
• Mécanique : Solide Indéformable from École Polytechnique Fédérale de Lausanne
• Mécanique Lagrangienne from École Polytechnique Fédérale de Lausanne
• Thermodynamique : applications from École Polytechnique Fédérale de Lausanne
• Planification et Design de Systèmes et Technologies d’Assainissement from École Polytechnique Fédérale de Lausanne
• Introduction aux Systèmes d’Information Géographique - Partie 2 from École Polytechnique Fédérale de Lausanne
• Global Arctic from École Polytechnique Fédérale de Lausanne
• Electric Industry Operations and Markets from Moscow Institute of Physics and Technology
• 農企業管理學 (Agribusiness Management) from National Taiwan University
• 普通物理學-電磁學、光學及近代物理 (General Physics (1)) from National Taiwan University
• 基礎光學一 (Introduction to Optics (1)) from National Taiwan University
• 材料力學一 (Mechanics of Materials (1)) from National Taiwan University
• Charles Darwin: El origen del evolucionismo moderno from Universidad Nacional Autónoma de México
• Una aproximación a la química del carbono from Universidad Nacional Autónoma de México
• Seguridad hídrica from Universidad Nacional Autónoma de México
• Introducción a la geotermia from Universidad Nacional Autónoma de México
• Introducción al estudio de los bienes comunes from Universidad Nacional Autónoma de México
• Conceptos y Herramientas para la Física Universitaria from Tecnológico de Monterrey
• Física: Dimensión y Movimiento from Tecnológico de Monterrey
• Mind of the Universe - Genetic Privacy: should we be concerned? from Leiden University
• Women in environmental biology from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• D'un infini à l'autre - Voyages de l'infiniment grand à l'infiniment petit from École Polytechnique
• Fundamentals of waves and vibrations from École Polytechnique
• Vers l'infiniment grand - Voyages de l'infiniment grand à l'infiniment petit from École Polytechnique
• Quantum Optics 2 - Two photons and more from École Polytechnique
• Density Functional Theory from École Polytechnique
• Vers l'infiniment petit - Voyages de l'infiniment grand à l'infiniment petit from École Polytechnique
• Les deux infinis et nous - Voyages de l'infiniment grand à l'infiniment petit from École Polytechnique
• Advanced Functional Ceramics from Yonsei University
• Tesoros de la Física y sus Descubridores I from Universidad de los Andes
• Arctic Development from University of Alberta
• Science Literacy from University of Alberta
• Getting started in cryo-EM from California Institute of Technology
• Fundamentals of Materials Science from Shanghai Jiao Tong University
• 生命科学发展史 from Shanghai Jiao Tong University
• Conceptos base para el estudio del medio ambiente from Universidad Austral
• Strategies for winning. Meteorology in a round the world regatta from University of Barcelona
• Técnicas Microscópicas de Caracterización from University of Barcelona
• Oceanografía: una clave para entender mejor nuestro mundo from University of Barcelona
• Water Supply and Sanitation Policy in Developing Countries Part 2: Developing Effective Interventions from University of Manchester
• Bajemos la temperatura: De la ciencia climática a la acción from Online Learning Campus - World Bank Group
• Large Marine Ecosystems: Assessment and Management from University of Cape Town
• Climate Adaptation in Africa from University of Cape Town
• Sex from Molecules to Elephants from Hebrew University of Jerusalem
• Metagenomics applied to surveillance of pathogens and antimicrobial resistance from Technical University of Denmark (DTU)
• Statistical Thermodynamics: Molecules to Machines from Carnegie Mellon University
• Ecosystems of California from University of California, Santa Cruz
• Introducción a los modelos de demanda de transporte from Pontificia Universidad Católica de Chile
• Deception Detox - using research methods and statistics to change the world from Erasmus University Rotterdam
• Planning for Climate Change in African Cities from Erasmus University Rotterdam
• Mind of the Universe: Science in Progress from Erasmus University Rotterdam
• 天文探秘 from Nanjing University
• 走进天文学 from Nanjing University
• Climate and Climate Change in New Jersey from American Museum of Natural History
• Our Earth's Future from American Museum of Natural History
• Fundamentals of particle accelerator technology (NPAP MOOC) from Lund University
• 有机化学 from Xi'an Jiaotong University
• Challenging Forensic Science: How Science Should Speak to Court from University of Lausanne
• La science forensique au tribunal: témoin digne de foi ? from University of Lausanne
• Comprendre l'écologie, pour une économie innovante from ESCP Europe

Humanities (197)

• A Life of Happiness and Fulfillment from Indian School of Business ★★★★★(508)
• Modern & Contemporary American Poetry (“ModPo”) from University of Pennsylvania ★★★★★(190)
• Miracles of Human Language: An Introduction to Linguistics from Leiden University ★★★★★(90)
• Introduction to Philosophy from University of Edinburgh ★★★★☆(75)
• Philosophy, Science and Religion: Religion and Science from University of Edinburgh ★★★★☆(60)
• Chinese for Beginners from Peking University ★★★★☆(59)
• Stanford Introduction to Food and Health from Stanford University ★★★★☆(46)
• Moralities of Everyday Life from Yale University ★★★★★(42)
• Child Nutrition and Cooking from Stanford University ★★★★☆(31)
• The Ancient Greeks from Wesleyan University ★★★★★(31)
• The Modern World, Part One: Global History from 1760 to 1910 from University of Virginia ★★★★★(25)
• The Emergence of the Modern Middle East - Part I from Tel Aviv University ★★★★★(25)
• English for Career Development from University of Pennsylvania ★★★★★(24)
• Philosophy, Science and Religion: Philosophy and Religion from University of Edinburgh ★★★★☆(23)
• Greek and Roman Mythology from University of Pennsylvania ★★★★★(21)
• The Modern and the Postmodern (Part 1) from Wesleyan University ★★★★★(21)
• English Composition I from Duke University ★★★☆☆(19)
• Sharpened Visions: A Poetry Workshop from California Institute of the Arts ★★★★☆(19)
• Patrick Henry: Forgotten Founder from University of Virginia ★★★★★(16)
• Magic in the Middle Ages from University of Barcelona ★★★★☆(16)
• English for Journalism from University of Pennsylvania ★★★★☆(15)
• Writing in the Sciences from Stanford University ★★★★★(14)
• Understanding China, 1700-2000: A Data Analytic Approach, Part 1 from The Hong Kong University of Science and Technology ★★★★★(13)
• Big History: Connecting Knowledge from Macquarie University ★★★★☆(13)
• Ancient Philosophy: Plato & His Predecessors from University of Pennsylvania ★★★★★(12)
• English for Science, Technology, Engineering, and Mathematics from University of Pennsylvania ★★★★★(12)
• Learn to Speak Korean 1 from Yonsei University ★★★★★(12)
• Søren Kierkegaard - Subjectivity, Irony and the Crisis of Modernity from University of Copenhagen ★★★★☆(12)
• Effective Altruism from Princeton University ★★★★☆(11)
• Plagues, Witches, and War: The Worlds of Historical Fiction from University of Virginia ★★★★☆(11)
• The Emergence of the Modern Middle East - Part II from Tel Aviv University ★★★★★(11)
• Constitutional Struggles in the Muslim World from University of Copenhagen ★★★★★(11)
• Intellectual Humility: Practice from University of Edinburgh ★★★★★(10)
• Age of Jefferson from University of Virginia ★★★★★(10)
• Ancient Philosophy: Aristotle and His Successors from University of Pennsylvania ★★★★☆(9)
• Philosophy and the Sciences: Introduction to the Philosophy of Physical Sciences from University of Edinburgh ★★★★★(9)
• Intellectual Humility: Science from University of Edinburgh ★★★★★(9)
• The Fall and Rise of Jerusalem from Tel Aviv University ★★★★☆(9)
• Soul Beliefs: Causes and Consequences - Unit 1: Historical Foundations from Rutgers University ★★★☆☆(9)
• Sports and Society from Duke University ★★★★★(8)
• Calvin - Histoire et réception d'une Réforme from University of Geneva ★★★★☆(8)
• English for Business and Entrepreneurship from University of Pennsylvania ★★★★★(7)
• More Chinese for Beginners from Peking University ★★★★★(7)
• The Holocaust - An Introduction (I): Nazi Germany: Ideology, The Jews and the World from Yad Vashem - The World Holocaust Remembrance Center ★★★★★(7)
• The History of Modern Israel - Part I: From an Idea to a State from Tel Aviv University ★★★★☆(7)
• Introduction to Ancient Egypt and Its Civilization from University of Pennsylvania ★★★★☆(6)
• The Modern World, Part Two: Global History since 1910 from University of Virginia ★★★★★(6)
• The Holocaust - An Introduction (II): The Final Solution from Yad Vashem - The World Holocaust Remembrance Center ★★★★★(6)
• Unethical Decision Making in Organizations from University of Lausanne ★★★★☆(6)
• Wonders of Ancient Egypt from University of Pennsylvania ★★★★★(5)
• Chinese Characters for beginner 汉字 from Peking University ★★★☆☆(5)
• Reason and Persuasion: Thinking Through Three Dialogues By Plato from National University of Singapore ★★★★☆(5)
• America Through Foreign Eyes from Rice University ★★★★★(4)
• 史記 (Shi Ji) from National Taiwan University ★★★★★(4)
• Étudier en France: French Intermediate course B1-B2 from École Polytechnique ★★★★★(4)
• Classics of Chinese Humanities: Guided Readings from The Chinese University of Hong Kong ★★★★☆(4)
• The Mediterranean, a Space of Exchange (from the Renaissance to Enlightenment) from University of Barcelona ★☆☆☆☆(4)
• A Journey through Western Christianity: from Persecuted Faith to Global Religion (200 - 1650) from Yale University ★★★★☆(3)
• Philosophy and the Sciences: Introduction to the Philosophy of Cognitive Sciences from University of Edinburgh ★★★★★(3)
• The Talmud: A Methodological Introduction from Northwestern University ★★★★☆(3)
• Chinese for HSK 4 from Peking University ★★★★☆(3)
• Understanding China, 1700-2000: A Data Analytic Approach, Part 2 from The Hong Kong University of Science and Technology ★★★★★(3)
• 中國古代歷史與人物：秦始皇（Qin Shi Huang) from National Taiwan University ★★★★☆(3)
• Pluralidades em Português Brasileiro from Universidade Estadual de Campinas ★★★★☆(3)
• Arch of Titus: Rome and the Menorah from Yeshiva University ★★★★★(3)
• Ancient Egypt from University of Pennsylvania ★★★★★(2)
• Understanding 9/11: Why 9/11 Happened & How Terrorism Affects Our World Today from Duke University ★★★☆☆(2)
• Know Thyself - The Value and Limits of Self-Knowledge: The Unconscious from University of Edinburgh ★★★★★(2)
• Sports Marketing from Northwestern University ★★★★☆(2)
• Chinese for HSK 5 from Peking University ★★★★★(2)
• 東坡詞 (Ci Poetry of Su Dong Po) from National Taiwan University ★★★★★(2)
• Corrección, estilo y variaciones de la lengua española from Universitat Autònoma de Barcelona (Autonomous University of Barcelona) ★★★★☆(2)
• El Valle de los Reyes from Universitat Autònoma de Barcelona (Autonomous University of Barcelona) ★★★★★(2)
• Gender, Family, and Social Change in Contemporary South Korea from Yonsei University ★★★★☆(2)
• Leer a Macondo: la obra de Gabriel García Márquez from Universidad de los Andes ★★★★★(2)
• The Science of Training Young Athletes from University of Florida ★★★★★(2)
• Big History - From the Big Bang until Today from University of Amsterdam ★★★☆☆(2)
• The History of Modern Israel - Part II: Challenges of Israel as a sovereign state from Tel Aviv University ★★★★★(2)
• Soul Beliefs: Causes and Consequences - Unit 3: How Does It All End? from Rutgers University ★★★★★(2)
• Writing your World: Finding yourself in the academic space from University of Cape Town ★★★★★(2)
• Cultural Competence - Aboriginal Sydney from The University of Sydney ★★★★★(2)
• Introduction to Who Wrote Shakespeare from Goldsmiths University of London ★★★★☆(2)
• The Holocaust: The Destruction of European Jewry from University of California, Santa Cruz ★★★★★(2)
• Russian History: from Lenin to Putin from University of California, Santa Cruz ★★★★★(2)
• Organising an Empire: The Assyrian Way from Ludwig-Maximilians-Universität München ★★★★★(2)
• Doping : Sports, Organizations and Sciences from University of Lausanne ★★★★☆(2)
• Data Science Ethics from University of Michigan ★★★★☆(1)
• Relativism from University of California, Irvine ★★★★☆(1)
• Lesson | Business English Skills: How to Navigate Tone, Formality, and Directness in Emails from University of Washington ★★★★★(1)
• 哲学导论（中文版） from University of Edinburgh ★★★☆☆(1)
• Philosophy, Science and Religion: Science and Philosophy from University of Edinburgh ★★★☆☆(1)
• Chinese Characters for beginner (2) 汉字(2) from Peking University ★★★★★(1)
• Le Bien, le Juste, l'Utile. Introduction aux éthiques philosophiques from University of Geneva ★★★★★(1)
• Deciphering Secrets: The Illuminated Manuscripts of Medieval Europe from University of Colorado System ★★★★★(1)
• 紅樓夢(The Red Chamber Dream) from National Taiwan University ★★★★★(1)
• 活用希臘哲學 (Understanding the Greek Philosophy) from National Taiwan University ★★★★☆(1)
• Learning Chinese : Start From Scratch (零到一學中文) from National Taiwan University ★★★★★(1)
• The Cosmopolitan Medieval Arabic World from Leiden University ★★★★★(1)
• ¿Cómo persuadir? Jugando con palabras, imágenes y números from Universitat Autònoma de Barcelona (Autonomous University of Barcelona) ★★★☆☆(1)
• La recherche documentaire from École Polytechnique ★★★☆☆(1)
• Re-imaging God in Korean Context from Yonsei University ★★★★☆(1)
• Intellectual Change in Early China: Warring States and Han from The Chinese University of Hong Kong ★★★★★(1)
• Historia de las reglas del fútbol en Inglaterra y en Argentina from Universidad Austral ★★★☆☆(1)
• Alle origini della civiltà mediterranea: archeologia della città dal Levante all’Occidente - III-I millennio a.C. from Sapienza University of Rome ★★★★★(1)
• Alfabetização Midiática, Informacional e Diálogo Intercultural - UNESCO e UNICAMP from Universidade Estadual de Campinas ★★★★★(1)
• Benjamin Franklin and His World from University of Pennsylvania
• Skepticism from University of California, Irvine
• Christianity & Ecology from Yale University
• Lesson | Business English Skills: Introducing Yourself in Business Settings from University of Washington
• Lesson | Business English Skills: How to Write Effective Openings and Closings to Emails from University of Washington
• Ukraine: History, Culture and Identities from University of Washington
• الأخلاقيات والثقافة ووجهات النظر العالمية from University of Illinois at Urbana-Champaign
• Éthique, culture et perspectives mondiales from University of Illinois at Urbana-Champaign
• Ética, cultura y perspectivas globales from University of Illinois at Urbana-Champaign
• Player Evaluation, Team Performance and Roster Management from State University of New York
• Magna Carta and its Legacies: Freedom and protest from University of London International Programmes
• Applied Public History: Places, People, Stories from University of London International Programmes
• Write Your First Novel from Michigan State University
• Luther and the West from Northwestern University
• Spanish for Successful Communication in Healthcare Settings from Rice University
• Essentials for English Speeches and Presentations 英语演讲与演示 from Peking University
• 悖论：思维的魔方 from Peking University
• 论文写作初阶（Academic Writing and Research） from Peking University
• 中级商务汉语（入职与营销篇） from Peking University
• 中级商务汉语——商务活动篇 from Peking University
• 现代汉语核心语法 from Peking University
• 中日文化交流史 from Peking University
• A l’avènement du christianisme : l’archéologie des derniers païens from University of Geneva
• Littérature mondiale from University of Geneva
• Langues et diversité: de la variation au plurilinguisme from University of Geneva
• Frontières en tous genres from University of Geneva
• Old Norse Mythology in the Sources from University of Colorado Boulder
• Leadership d'excellence par le sport de haut niveau from ESSEC Business School
• 莊子─人情 (Zhuangzi─Between People) from National Taiwan University
• 是誰在說話 -- 可愛的臺灣。（Intermediate Chinese） from National Taiwan University
• 現代文學導讀：詩、散文、小說 from National Taiwan University
• 又見大觀：《紅樓夢》中的烏托邦 (The Red Chamber Dream (3): Daguan Garden) from National Taiwan University
• 深扣孔孟 (An Inquiry into Confucius and Mencius) from National Taiwan University
• 中國園林文學 (一) (Chinese Garden Literature (1)) from National Taiwan University
• 紅樓夢-母神崇拜 (The Red Chamber Dream (2)) from National Taiwan University
• 東亞儒學：孟子一(East Asian Confucianisms: Mencius (1)) from National Taiwan University
• 東亞儒學：孟子二 (East Asian Confucianisms: Mencius(2)) from National Taiwan University
• 唐詩新思路 (Tang Poetry) from National Taiwan University
• 東亞儒家：人文精神一(East Asian Confucianisms: Humanism (1)) from National Taiwan University
• 發展心理學：哲學觀與方法論 (Developmental Psychology: Philosophical Bases and Methodology) from National Taiwan University
• 東亞儒家：人文精神二(East Asian Confucianisms: Humanism (2)) from National Taiwan University
• Química, guerra y ética from Universidad Nacional Autónoma de México
• Fundamentos de la escritura from Tecnológico de Monterrey
• Fundamentos de la escritura académica from Tecnológico de Monterrey
• The Importance and Power of Music in our Society from Leiden University
• The Rooseveltian Century from Leiden University
• Unión Europea: Historia, Instituciones y Políticas from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Oralpha: Literacy for new citizens, a course for trainers from Universitat Autònoma de Barcelona (Autonomous University of Barcelona)
• Leer a Macondo: la obra de Gabriel García Márquez from Universidad de los Andes
• Try RRI! A guide for Responsible Research and Innovation from University of Amsterdam
• Media ethics & governance from University of Amsterdam
• Training and Practicing in English Public Speaking from Shanghai Jiao Tong University
• 中医药与中华传统文化 from Shanghai Jiao Tong University
• Structuring Values in Modern China from The Chinese University of Hong Kong
• Religion and Thought in Modern China: the Song, Jin, and Yuan from The Chinese University of Hong Kong
• La estrategia china from The Chinese University of Hong Kong
• Religious Transformation in Early China: the Period of Division from The Chinese University of Hong Kong
• Conciencia Lingüística from Universidad Autónoma de Madrid
• Let’s Read! Learning Japanese through Science & Technology-1 from University of Tokyo
• Let’s Read! Learning Japanese through Science & Technology-2 from University of Tokyo
• I. Missões jesuíticas nos territórios guaranis: história from Universidade de São Paulo
• 《新教伦理与资本主义精神》导读 from Fudan University
• A Voice of Their Own. Women's Spirituality in the Middle Ages. from University of Barcelona
• American Deaf Culture from University of Houston System
• Modern Hebrew Poetry שירה עברית מודרנית from Hebrew University of Jerusalem
• Moses' Face: Moses' images as reflected in Jewish literature from Hebrew University of Jerusalem
• AI, Empathy & Ethics from University of California, Santa Cruz
• Odessa: Jews in the Modern World from University of California, Santa Cruz
• Odesa: Jews in the Modern World from University of California, Santa Cruz
• The History of Medicine: Philosophy, Science, and Psychology from University of California, Santa Cruz
• Asian Environmental Humanities: Landscapes in Transition from University of Zurich
• Introducción a la Ingeniería Gastronómica - ¡Ciencia en la Cocina! from Pontificia Universidad Católica de Chile
• Thought Experiments: An introduction to philosophy from Erasmus University Rotterdam
• Reality Bites: Introduction to metaphysics from Erasmus University Rotterdam
• The Epistemic Quest for Truth: Introduction to epistemology from Erasmus University Rotterdam
• A Guide to Diversifying and Transforming Football Leadership from Erasmus University Rotterdam
• The Politics of Skepticism from Erasmus University Rotterdam
• 理解马克思 from Nanjing University
• L’Egitto prima e dopo i faraoni. Dalla nascita dello Stato (3000 a.C. ca.) alle prime comunità cristiane (IV secolo d.C.) from Sapienza University of Rome
• Egypt before and after pharaohs. from Sapienza University of Rome
• Paesaggi di Roma Antica. Archeologia e storia del Palatino. from Sapienza University of Rome
• Writing in English at University from Lund University
• Akademiskt skrivande from Lund University
• Unravelling solutions for Future Food problems from Utrecht University
• Improving Your Statistical Questions from Eindhoven University of Technology
• 中国哲学经典著作导读 from Xi'an Jiaotong University
• 《论语》的智慧 from Xi'an Jiaotong University
• 品读道家的智慧 from Xi'an Jiaotong University
• Dopage : Sports, Organisations et Sciences from University of Lausanne
• Science and Technology in the Silla Cultural Heritage from Pohang University of Science and Technology
• Chosen Issues in Holocaust History from Yad Vashem - The World Holocaust Remembrance Center
• El Holocausto: el abismo de la humanidad from Yad Vashem - The World Holocaust Remembrance Center