I am currently a postdoc and my PhD was in applied mathematics in the area of numerical analysis and electromagnetic/acoustic wave propagation. There was no statistical element to my PhD, it was completely deterministic. I took several probability/statistics and one machine learning module 5-6 years ago during my BSc, and a stochastic ODE module during my MSc but that's about it..its been all applied mathematics since then.

I am considering leaving academia and entering industry and it seems like there are far more jobs in the area of data science/machine learning than there are for my skillset.

  1. If I left academia and began 'studying up', how long do you think it could take me to gain the skills required for a data science/machine learning position in industry?
  2. It seems like there is a very wide variety of science/machine learning techniques and obviously there isn't time to learn all or even most of them. So what approaches are absolutely essential for data science/machine learning in industry these days and what is the most efficient route to gaining these skills?

As the market is in desperate need of people, and there are plenty of people with absolutely no formal training and no background in statistics, you are already perfectly qualified to spin this hype wheel and call yourself a "data scientist", too.

I'm not kidding. Just do some free online courses and you'll likely see that you can do all they ask for. Data science is about buzzword bingo, not about being smart at statistics not good at coding (unfortunately).

If you don't want to feel like an impostor, I suggest the following: find some important algorithm still missing from the big toolkits such as sklearn, R, Weka, ELKI. Implement it, and contribute it to some open-source toolkit. Then you can call yourself an "sklearn contributor" in your resume, which puts you ahead of 90% of self-proclaimed data scientists. What could make you a more proven data scientist / machine learner than having written code used by other data scientists / machine learners?

  • $\begingroup$ Contributing to an open-source toolkit sounds like a good idea alright, I will certainly look into that. $\endgroup$ – electroscience Apr 1 '19 at 7:04

I think you already know enough applied mathematics to begin with. You can pick-up rest of it as required.

One option is :

  1. Start with an online course that provides high level overview of machine learning and types of algorithms (E.g.: https://www.coursera.org/learn/machine-learning)
  2. Start applying the knowledge in real world problems as soon as possible.
  3. Learn various types of neural networks (deeplearning.ai is one place to start)
  4. Apply knowledge to real world problems (Such as Audio/Video classification , Natural Language)
  5. Get an internship

This will take 5 - 6 months.

  • $\begingroup$ Thanks, I've just enrolled in some of those courses. $\endgroup$ – electroscience Apr 1 '19 at 7:25

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.