meme

courtesy

TLDR

  • I came to know some people are interested in AI/ Machine Learning (ML) but are from non-CS background
  • Solution is simple: Learn, Build and Share
  • There are some popular pathways for Machine learning, each of them have pros and cons.

Background

For the last few days, I see a trend. Some Non-CS University juniors asked about career / higher study in Machine learning / AI in a facebook group of my university. Nice to know 🙂. I think many other people have the same questions but they are shy/afraid to post (hence explaining anonymity in the group posts maybe) about this field. I think I should share some guidelines from my experience. I have been working as an AI Engineer for 4 years. I have worked on developing proof of concept ML models, Direct software based on AI for robots, and now working in a legal tech company. Different people approach learning AI / ML differently, I will share mine. You have to note that, this worked for me, but may not necessarily work for you. You have to adjust according to your schedule/lifestyle.

But the basic principle is the same: Learn, Build, and Share. It means you keep learning new stuff every day, Build some solution using it, then must Share what you did.

Courses to start the journey:

  • 1) Machine learning Specialization by Andrew Ng
  • 2) Deep learning specialization by the same guy
  • 3) course.fast.ai
  • 4) workera.ai
  • 5) End-to-End project tutorials from Youtube.

Approaches to learn ML by yourself.

There are several approaches one can take.

Explore and Solve Problems in Kaggle

Pros

  • It will give you a good idea about data science and machine learning.

Cons

  • But there is one problem, most real-world problems are not like Kaggle. You have to do a lot of preprocessing steps to adjust the data and understand it.

Try to make some software using data science and ML

Pros

  • Python is more than enough. It will help you tackle real-world problems using Machine learning and also will give you some experience in software engineering.

Cons

  • The issue here is you will have to deal with many problems which are not directly ML related but are very important for the applications. And many times this process is boring. You may get stuck in one type of problem. You have to balance it with learning new stuff.

Keep a schedule for learning new stuff and/or honing basic skills

Pros

  • The issue is AI is a very fast-changing field. If you get stuck with one type of architecture, you may lag behind others.. You need to keep updated. ( i do it myself) . It can be watching videos every day, reading papers every day, or doing some extra projects every day.

Cons

  • Nothing . Learning is never a bad thing.

Bonus steps but very helpful for getting jobs

  • Keep pushing in GitHub whatever coding you are doing. Engineers judge others based on Github projects.
  • If it is possible try to contribute to open source projects on GitHub. ( I myself ain’t able to do that yet due to time schedule).
  • You can test your skills in workera.ai. They are good for guiding people in their AI careers. [Note: I am not sponsored by workera]

Changing Major

I think some students are afraid/shy of changing their Major. “AI is a CS-only subject” and bla bla bla. I ask you to stop treating your Majors as Religions. Just because you are not a “Pure” Mechanical Engineer / Civil Engineer, it doesn’t mean your life is gone. Learning and working on AI will not only help you open new career prospects, it will also give you higher study prospects in your own department, as many new types of research are being conducted in non-cs subjects using ML. There are many things I can share, but this post is already too long. Don’t want to waste time. please feel free to message me anytime.

Sincerely, Your mechanical engineering brother with a very low CGPA 🙂