I actually went to that iMechanica site and got in contact with a professor there looking for PhD candidates. Unfortunately, my GPA was waaaaaaaaaay too low to ever be let into their school, especially as a PhD candidate without a Master's.
But I did learn what's marketable in the work that I want to do and it's physics, GPGPU computing and, ironically enough, web development.
This actually inspired me to pick up CUDA again and now that I've had time and about 8 months of professional experience in something I knew relatively little about, I find myself having a much easier time. Hopefully, I'll be able to actually be the first person to write a "degenerate" GPU-based version of a Delaunay triangulation that isn't horrendously slow. At least, I think I would be the first. I've tried googling this so much that if there's someone out there, tell me!!!!