1
$\begingroup$

I am a graduate student in Mathematics, and I have been recently offered a post-doc at a good university. While it is not a top school in the most strict sense of the word, it is one of the best positions in my field of research.

However, I have recently started wondering whether academia is really the right choice for me, and I am considering learning data science with the intention of transitioning towards industry.

The question is, assuming I was to take that step and decide to start studying towards leaving academia, would it be better for me to take the post-doc, or perhaps stay one more year in graduate school?

The obvious advantage of taking the post-doc is that it would give me ample time of financial stability to learn new things. However, I am mostly interested in an answer from the point of view of starting a career in data science: is a post-doc in Mathematics something that would be considered a worthwhile addition to one's resume, or does that perhaps make one look a little bit too "academic" and actually impede one's ability to get an industry job?

The advantage of staying for one more year in graduate school would be perhaps be that since it is a top American program, I would have access to plenty of resources that I am not sure would be available at a good European university. Moreover, it seems to me that there are more good opporunities in data science in the US rather than abroad. Thus, if my goal was to get a job here, I wonder if it wouldn't be easier to just not leave the US in the first place..? I am not a US citizen, and so work authorization is something I have to consider.

$\endgroup$
  • $\begingroup$ While I can't really answer your questions I can add a little something. I've worked with mathematical PhDs and people from math. While they have great theoretical knowledge, they lack the understanding of how to turn it into value and products. Some lack practical skills such as software development, tooling, and practices which makes them hard to hire. $\endgroup$ – Carl Rynegardh Feb 23 at 20:04
  • 1
    $\begingroup$ According to this post veekaybee.github.io/2019/02/13/data-science-is-different the US labor market for data scientists is oversaturated. Besides from this, a Math Postdoc helps particularly if it is in an applied field. $\endgroup$ – knb Feb 23 at 21:33
  • $\begingroup$ @knb Clearly you have not tried to hire a data scientist recently :-). While there are data scientist resumes available out there, most of them are plain awful and don't know what they are doing. A good data scientist is still extremely difficult to find $\endgroup$ – I_Play_With_Data Feb 24 at 1:29
1
$\begingroup$

There was perhaps a time where a math Phd would be an immediate ticket to a data science position but I think that phase has come and gone. The reason? It has a simple answer: Tensorflow.

Tensorflow is essentially all of Google's Phds getting together, writing out the formulas for a large number of data science models and, in the process, that takes a lot of the math out of data science. And this makes sense since most math in data science is "write one code" anyways.

In a similar vein, DataRobot is effectively the best Kaggle competitors in the world getting together and also taking a lot of the math out of data science.

So, I hate to be the bearer of bad news, but the math Phd just isn't attractive to most data science teams.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.