I have a couple of years kf experience as SDE and I’m working towards my part time masters in ML. I find it very difficult to get into ML roles as almost no one hires inexperienced ML folks. How to break into it? Is doing a full time masters the only way?
Hello, I have been in your shoes a couple of years ago. I was working as BI engineer in Silicon Valley and realized the potential of ML and Data Science. So I switched gears towards learning algorithms. I learned from logistic all the way to Keras. But as you said, everything is theoretical. As I wasn't working on the practical part of implementing, it has been difficult for me to crack the interviews as most of the questions are related to the project. They ask about your current project in ML and go deep dive in asking why you haven't used a different algorithm, etc etc. I gave interviews and failed in every one of them. But the good thing is that, I have learned through the process and cracked one finally. it is still a mix of BI and ML. but no complains as I don't have handson experience in ML. I suggest you to pick up python, as almost every company asked me to code an algorithm. As you are already an SDE, it shouldn't be of big concern. And the main reason is that, python has vast libraries for data science. Go through the underlying definitions of all ML algorithms. Pick a ML project which your company is working and try to get as much exposure as possible. Trust me, the part-time masters doesn't really matter in cracking a job. It all depends on how well you explained the project, why you implemented a specific algorithm and not the others, challenges in data analysis before implementing ML and of course, coding interview. Good luck.
Please give a try to <a href="http://www.fast.ai" rel="nofollow">http://www.fast.ai</a> free courses on practical machine learning. These courses had huge impact on people wanting to enter into the world of machine learning like me.
Several ways to get noticed in the application process:
* Go to Kaggle and participate in some of their competitions. * Build an ML program from the ground up to help build your portfolio (something like a dog door that only opens to your dog).
* Implement a paper from arxiv - try to understand what the paper is about, then download the data set and implement the model they are talking about and see if you get similar results. Then, write a blog post about what you did....
It may be easier to get a position as a more general purpose engineer (full stack or whatever you’re most comfortable in) with a team that has ML engineers, and work your way into the role you want from there.