I've noticed that ads for postdoctoral positions emphasize the skill set that one must have for a particular position. That said, what are the areas of research to avoid because they give you few transferable skills and hence limit your range of possible postdoctoral positions?

For example, String Theory might be an area to avoid because it gives you virtually no experience with using scientific software packages that might essential for post docs in other areas. Any others?

(My motivation for asking this question is that I think it might be nice to roam into other interdisciplinary areas and fields post-PhD, rather than restrict oneself to a certain field for life.)

  • $\begingroup$ There are no research areas with less transferable skills, as to get good at one, you have to know a lot of others too. Also, vote to close, as make a list, and a nasty one at that. $\endgroup$ – Ron Maimon Sep 17 '12 at 9:56

Coming from a computer science standpoint, I don't know so much about which fields to avoid.

However, I will highly encourage serious physicists to take up general GPU programming. On my side, the boon of physics is obvious. However on research especially dealing with sensors and data, there are so many applications we will likely not run out.

We (the GPU programming field) need more cross-discipline understanding, especially from the areas of physics, molecular biology and genetics. It can be a little difficult to explain the full need of having a deeper 1 to 1 relationship in GPU based tools.

I will offer the only practical advise I have thought of in this area: anything without some connection to current material science (carbon nanotubes, advanced magnetics, quantum processing, bioprocessing/storage, advanced ceramics, etc) seems to have difficulty finding the placement it sometimes deserves in current trends.

For instance though there's great need for advancements, things like geo-imaging and medical imaging are nicely thriving. Tools are made for all sorts of spectroscopy, while some areas of physics are not so applicable.

Meanwhile, I'm hoping for more cross-talk in our fields, because we would very much like to make tools in exploration of the more exotic realms of physics.

  • $\begingroup$ This is reasonable advice, but physicists generally all have to know how to program, in all subfields. The GPU thing is not likely to be of permanent interest, as the GPU is just a hack to get better parallelism out of existing hardware. It will probably lead to next generation hardware integrating GPU and processor more. $\endgroup$ – Ron Maimon Sep 17 '12 at 9:53

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