AI/ Deep Learning/ Machine Learning

AI/ Deep Learning/ Machine Learning
0

#1

I finished all the challenges on here about a month and a half back. Since then I’ve worked on my own projects, cloned a couple. I really enjoy designing front end and writing backend equally. So I knew I would love to be a full-stack engineer.
But lately, something else has been grabbing my attention. AI/Deep Learning/Machine Learning.
I love everything about it. It’s really cool to see the demos people upload, how they train their own neural networks.
So I guess my question really is, is all the hype just hype or are there real opportunities in this field for everyone? What kind of salary(I understand it varies with the role, but what range of salary) can one expect?
If I was to start learning Machine Learning/Deep Learning, how would you suggest I get started?
Any input is much appreciated. Thank you!


#2

There is always room in this field. I can’t speak to salaries, I’ve not done much research into the money, but if you look around on sites like glassdoor or see if you can find a job description that includes salary ranges you’ll get an idea. Unless you live in a city with an unbelievably high cost of living, I would expect that there is no position in ML/DL/AI that doesn’t pay a living wage and then some, so you’d do yourself best to find something you’ll enjoy doing.

Personally, I suggest that people start looking into the different roles and responsibilities within the field. Try out a little bit of everything and see what you enjoy doing most. I would have never thought I like data engineering and architecture, but after a little experimentation it’s become my favorite part of the whole process.

JavaScript is getting a number of good resources to work with data, but languages like Python and R are more industry standard (though standards do vary within different areas within the field, like I mention here). So, if you want to stick with JS, there is a niche for that, otherwise a good place to start is to find out what you like to do, and find out how everyone else is doing it (or make a new set of standards to work by, do what makes you happy).


#3

Thank you.
After reading your input, I did a search on Linkedin and, as you said, there are jobs in this field but they almost always need at least 3 to even 10 years of experience.
After a little more research, I came to the conclusion that, one must start as a Data Scientist/ Data Engineer.
Am I right? If yes, could you please answer this question.


Thanks a lot!


#4

After a little more research, I came to the conclusion that, one must start as a Data Scientist/ Data Engineer.
Am I right?

Not at all. Data Science is a broad umbrella term that includes fields like Machine Learning, and Data Engineering has it’s place, but is not machine learning. If you want to get into machine learning, you need to be doing machine learning.

I don’t know anyone who got into it from following a sequential curriculum. You’ll need to make sure your math and basic programming skills are on par, otherwise, everyone I’ve ever talked to found a project they wanted to work on and just went for it, filling in knowledge gaps as they went along.

Don’t worry about experience required, no position is ever filled by a candidate who meets every single requirement because in addition to the skills they have listed, they (more importantly) need a person who can work on and grow with their team. Job listings are meant to describe an ideal candidate, not a realistic one.


#5

If my end goal is ML, would starting out with Data Science, instead of jumping right into learning Math needed for ML, be a waste of time?
Also, where to start learning ML? What Math concepts should I be crystal clear on? Where can I learn them?
Some people directly start with libraries like TensorFlow. Would it be worth it to learn everything from scratch? I hear that if you wish to write your own neural network or understand things from the ground-up better, then you must learn from scratch( I mean all the Math involved).
Anything you can tell me will be very helpful. Its really unfortunate none of my classmates or professors in college are passionate about anything. So your help is very much appreciated. Thanks a lot!


#6

Like I said in my last response, Data Science is a broad umbrella term that includes machine learning. Asking if starting with data science instead of machine learning is a waste of time is a lot like asking if starting with programming instead of back-end development is a waste of time. Machine Learning is a part of data science just like back end development is a part of programming: It’s not always required, but in the instances where it is, it’s best to have a specialist do it than someone who only kind of knows what’s going on.

It is possible to take the black box approach and just dive into frameworks that do all the hard work for you but, increasingly, employers are turning down candidates who do so (and they are right to). With the increasing concerns of ethics in machine learning and AI, it is extremely important that developers are able to explain what their algorithms are actually doing and are able to quickly and effectively adjust input features in the event of biased outputs (garbage in, garbage out). To my knowledge, the only math fundamentals you need down are calculus, linear algebra, and basic statistics.
If you need a slower, more in depth resource, I love Khan Academy. They go too slow for my attention in most circumstances, but when I Can stick with it I never have questions about what he was explaining. If your math skills are already pretty solid, Microsoft has a kind of overview of the essential mathematics (They don’t explain concepts thoroughly, so I recommend having a good mathematical foundation before going in). The Microsoft course is part of a larger AI curriculum, but I can’t speak to the quality of any of the courses that come after the essential math one, and the Python one felt like it covered the topics it did well, but that a bunch of important topics weren’t covered at all. So you’d have to supplement your learning, but if you want to give that curriculum a try, it probably won’t hurt.


#7

In response to this, I just came across this blog post on “Learning Math for Machine Learning”. It quotes some solid review material for linear algebra (PDF) and probability theory (PDF) that may be needed for machine learning.

The biggest thing on math is just to believe you can learn math.

The truth is, people who are good at math have lots of practice doing math. As a result, they’re comfortable being stuck while doing math. A student’s mindset, as opposed to innate ability, is the primary predictor of one’s ability to learn math (as shown by recent studies).

Good luck, @graven_whismas!