I have an EECS background but have spent most of my career in PM/UX roles. I haven't regularly programmed in about 10 years. Most of my projects in that time have involved light web prototyping and quickly hacking things together for demos. It's fair to say I couldn't pass your average whiteboard or leetcode interview at this point.

I've started to lose passion for my role (and company) and am interested in building things again, likely at a small company or startup. After a lot of dabbling, I've decided that I'd like to transition to being an ML Engineer and I'm looking for advice on how best to make that happen. My questions are:

(1) What educational resources would you recommend, both in terms of helping me gain knowledge and improving my odds of getting a job? I've been working through Fast.ai, but am not sure if I should pursue others (Coursera, or even a Masters or Grad Certificate of some kind).

(2) For MLE interviews/roles, will I need to display strong theoretical strength (i.e. in advanced statistics, ML foundations) or will applied knowledge and portfolio projects be enough?

(3) What strategy would you recommend for someone in my shoes? Fill a GitHub profile with projects and aim for entry level startup jobs? I'm willing to "start at the bottom" but would hope that my experience counts for something. Would any company even hire someone like me as an MLE?

I'd love any input to the above, but if you have general thoughts to share about someone in this situation that would also be helpful. Thanks.