PhD for Data Science/Machine Learning

Part 1 (2018)
1

Hi,
I recently attended a meetup on Data Science and noticed that all the panelist were PhD holders (not necessarily in mathematics or computers science with the exception of 1 panelist). I also notice that both Jeremy and Rachael are PhD holders. I am wondering is this a pre-requisite to enter the field of Data Science/Machine Learning. Are they any Data Scientist/Machine Learning Engineers who are renowned and great practitioners not necessarily holding a PhD.

My initial interest in this area arose from a problem that I encountered at a customer and on further digging I discovered ML / Deep Learning & Data Science.

Regards
Murali

2

Jeremy’s LinkedIn page says he has a BA in Philosophy, but I don’t see an entry for a PhD.

I recommend these fast.ai posts:

Machine learning hasn’t been commoditized yet, but that doesn’t mean you need a PhD

Alternatives to a Degree to Prove Yourself in Deep Learning

To become a data scientist, focus on coding

These posts will link to more posts related to your question.

3 Likes
3

@Matthew I guess I got that wrong about Jeremy.
I will read through them.

4

There are some major advantages to holding a PhD. I don’t have one, but I was in phd program for almost 3 years before I dropped out. The big skill you learn is how to read papers and apply new ideas to solve a novel problem of importance. You can think of a PhD as experience working on a project where there is both conceptual and technical activity, that somehow move a field forward through publication.

That said, I work in industry leading the development of machine learning algorithm for a start up company (Data Science), and not having a PhD doesn’t hurt me per se, my area was computational biology anyway. I’ve worked with extremely skilled people with and without a PhD, it all comes down to their drive and knack for accumulating relevant experience.

If you want to do Deep Learning, then do Deep Learning. Commit to it. For me, this is going to be one approach in a tool belt of many options, but I would look very favorably on job applicants capable of really grokking differential programming like PyTorch.

1 Like
5

@wespiser thanks for your insights. when I did attend the session couple of days back it appeared a bit demotivating when there was so much emphasis on PhD

My initial foray / interest into ML/Deep Learning started from a problem that I encountered and was left wondering why should I do this why cannot a machine do this task as we have the necessary data.

@jeremy @rachel your thoughts would be highly appreciated