What do I think about Data Scientist
T-Shaped People have deep expertise in a specialist field, coupled with a broad skillset, which allows them to work and collaborate outside their area of expertise.
You need to build your knowledge as T-shape, horizontally, financial knowledge, hardware knowledge; at least you need to know enough to communicate with the experts in the field, so you can apply your own expertise to it. Vertically that is your expertise, for example, statistical machine learning, nonparametric statistics (e.g. non-parametric Bayesian Hidden Markov Model), …etc.
Horizontally, you need to pick up one or two the required undergraduate or graduate courses from each disciplines related to the application you want to go to (let’s say, medical images, electronic manufacturing, investment strategy, consumer e-commerce etc.) Vertically, you need to master the graduate school courses in your field and be able to keep up with a few recent advanced papers in your field (reproduced and extend to new fields or new applications).
School prepare you well for general knowledge, but in work place to solve specific problems you will need a lot of “know-how”(which most of time is not your specialty) to solve the specific problem your boss assign to you.
You need to build your competitiveness on certain domain knowledge; machine learning deep learning skills are just tools.
In general, project management skills and product management skills are always needed in work place. Winning research proposal writing skills is a must for academic career.