Tech IndustrySep 20, 2018
MicrosoftAqfN57

Can you compete as a machine learning engineer without a masters degree?

Title says it all

Intel Obsvr_J Sep 20, 2018

In what competition? If as in Applied ML I think you have a good chance. If as in job search you might find it harder IMO

Oracle ovKR85 Sep 20, 2018

Absolutely, in any field.

Zillow Group sofaking Sep 20, 2018

Yes. The author of pytorch-nlp just graduated from college. Now he’s speaking at NLP conferences. He’s also wicked smart. That said, there are a lot of ass-backwards corporations out there who won’t even look at your resume because their systems filter out applicants who don’t have a post-graduate degree. They are not necessarily wrong, but that means if you happen to be one of the top 5% of outliers like the person I referenced to above, you’re playing against the system. Then again, at those levels you don’t really need the system to work for you because you’d be able to find your own path easily.

Uber cfsc Sep 20, 2018

They’re not necessarily ass-backwards. As with any reasoning and decision-making in real life, you take a Bayesian approach and allocate your limited resources. It may be that a farm boy from a remote mountain village invented deep learning and built a prototype GPU to run it on in his spare time with straws, but the chance is not very high. Hence, it does not make sense to call him in for an interview, even given the nonzero chance he’s got the potential to be the creator of SkyNet.

New
uNSi04 Sep 20, 2018

Talk directly to the people in research labs. Recruiters generally don't get that you can do ML work without a PhD

Uber cfsc Sep 20, 2018

It’s not just recruiters. It’s simply that the likelihood of you being able to do cutting edge research is higher given that you have a PhD than otherwise. If you can do some side projects and give additional observations to condition on (like Kaggle competitions, independent publications, built self driving car in garage), then I’m sure the impressions will change.

New
uNSi04 Sep 20, 2018

I'm finding that even those side projects have a much higher impact when presented to a researcher than a recruiter. A researcher knows that you've got what it takes even by listening to your idea. Recruiters will look for more traditional metrics such as "achieved ... % accuracy...". You probably know better than me in this area but as a new grad this is what the kind of response that I'm getting on my applications.

PepsiCo akaGhost Sep 20, 2018

Rule of thumb when you’re self-taught: at some point you have to stop learning and start building. I don’t care if it’s a website that matches flowers or an app with three users, you don’t have to demonstrate your creativity or entrepreneurial chops, just your mastery of the skills. Just put down the tutorials and textbooks and build a system or three that solves a problem, even if that problem is stupid or silly or frivolous.

Microsoft DOyu71 Sep 20, 2018

The ones I know, maybe 1 in 5 has a higher level degree.