[numerik_meister]

Early on in my career I noticed that the most fun programming seems to be done by people who are not actually trained as computer scientists: mathematicians, physicists, engineers, chemists, astronomers ...etc This sort of numerical computing and simulation stuff is incredibly fun and quite rewarding. You work much more often on core logic and much less often on boring fluff. Of course it is all very subjective and it pays shit. So the answer is: research at a research uni, employed by an applied mechanical engineering lab, but trained as a mathematician.

[fxtentacle]

Agree, I studied mathematics and molecular biology and the most interesting problems are intersections between that and coding (which I learned by myself as a kid).

[mvind]

Have you experienced a caveat in terms of coding quality? I study mathematics, and I think it is pretty standard for mathematicians to disregard theoretical run times when experimenting and doing "napkin" computations. Invariably this leads to relatively poor code quality. Speaking from experience, I just failed a coding interview because I solved the question like a mathematician would do, i.e. a quick and dirty way. What is your experience with this phenomenon? I know academics often gets railed lacking concepts in data structures etc.

[tgamblin]

You can do this kind of work and also get paid decently at a national lab. For example, LLNL is hiring!