How difficult is it for someone with a MSCS to get a MLE role if didn’t do the degree in ML and have no professional ML experience? I have around 3 YoE as a SWE, in NYC, and this is assuming I taught myself (online courses, YouTube, projects, etc) the required info listed on job listings. Thinking of the rating as a combo of actually being qualified for the role, getting an interview, getting through the interview, overall competition/crowdedness, etc Side question not for the poll, is a PhD really needed for the cool stuff in ML?
No, you can go MLE/DS without PhD, PhD is only needed for RS.
PhD only if you want to work in research, it's very rare that people without PhD are given research responsibilities. For engineering, you just need to be a good swe with knowledge of ML systems.
What do you mean by cool stuff in ml? Either you are competent enough to do the job or not? Are you willing to go to heart surgery to a dr who learn surgery by watching YouTube videos?
that's a bad example. online learning is monumental in software engineering with low cost of failure as opposed to learning surgery online and killing patients. research is difficult to learn online (requires discipline) but online learning is very well suited for software engineering
I think people are talking about ml/ai here
It's doable. You can probably transition gradually, if your team has ML work. Just start asking your manager, if they can give you some ML related projects. Small at first, and bigger later. I'm MLE at a startup. In general, I like it, but I got tired of the way ML is organized in my current company (we're in "magic" AutoML space). So I find, I'm much happier, when I don't need to tweak models for the customers again and again, but to work on implementation/engineering tasks. I imagine though, that working a bigger company, in which ML is serving themselves, has it's own established pipelines and working ML models would be very different. I started considering switching my jobs for that.
Most people who end up in MLE end up disappointed and jaded, since the day to day tasks of MLEs are identical to that of SWEs. The people who are actually doing interesting things are usually classified as RS and usually have PhDs.
Just go work on ML Infra/Ops teams. It’s just SWE work anyway.
Do you specifically need ML experience for it or is “normal” SWE work considered good experience?
Yes, you do need ML experience for the interview.