Right Qualifications for DFT in Industry

[maverick_starstrider]

Hi All,

I'm finishing up my PhD in Computational Condensed Matter. I've worked almost exclusively in spin systems using things like Monte Carlo and Exact Diagonalization. However, my specialty is in producing high-quality, high-performance, often parallel research codes (i.e. none of my codes have used other libraries, they're all written by me and probably internationally competitive for the exact esoteric problem I was working on). I've also had 3 industry co-op jobs utilizing a variety of numerical techniques.

So long story short I'm a condensed matterist with a pretty decent and robust background in computational theory and HPC (high-performance computing).

I'm now doing the job hunt and what I'd like to do, in a perfect world, is do simulation/numerical work on sort of nanoelectronic devices such as spintronics, graphene, crazy heterojunctions, etc. However, I'm starting to get the impression that most of this industry or near-industry postdoc work is using DFT (Density Functional Theory) and seems to just be done using available open source software libraries.

I have no experimental skills and I don't know how attractive my HPC background would look if these groups are just using some black box libraries.

Do you guys think these sorts of jobs are a slim chance for me?

 

[cgk]

If you are looking for a post-doc, I would not see any problem with this. Doing something different than you did before is the point of a post-doc, after all.

But just a comment on the DFT & similar approaches: I would recommend not calling them "black-box libraries". And the reason people "just use" them instead of making their own is that making a competitive solid state DFT program is *at least* two orders of magnitude more complicated than hacking up some Fermion exact diagonalization or quantum Monte Carlo program. Nevertheless, if you would like to focus on the development of such programs: That is also a scientific discipline, followed in many groups (e.g., look up Georg Kresse in Vienna, or Juerg Hutter or Joost VandeVondele in Zuerich). Typically you would then work on some aspect of improved algorithms, and do some applications that profit from this.

 

[maverick_starstrider]

Thanks so much man. I actually looked into the groups (and others) that you posted. I appreciate the reply.