What are interesting topics to solve in astrophysics for computer science graduate?

One interesting problem that was posed to me by my former advisor was a new type of mesh that alleviated the problems in both grid-based (read: density-based) simulations and SPH (smooth-particle hydrodynamics, basically just the particle Lagrangian method) that used a tesselation of Voronoi shapes to make up the grid. Here is the paper describing the method: http://arxiv.org/abs/1109.2218 For the OP in particular, this doesn't seem like a great first problem to work on because you're building from the ground up (which is a massive, massive undertaking in itself.. codes come from theses, usually, which is a lot of full-time work). A simpler thing to do would be to get familiar with these concepts: Finite difference, finite element, finite volume, spectral, multigrid, and other methods for solving nonlinear partial differential equations Probably a good idea to know ODEs/PDEs and nonlinearity quite well (for nonlinearity, try Strogatz's book "Nonlinear dynamics and chaos", should be simple for someone with some background in ODE/PDE, physics, and/or linear algebra) Astrophysics (depends on what scale you're looking at, stellar diffusion? galactic dynamics? distribution of dark matter in the universe?) If you don't care to learn or don't have the time to for all of that stuff, you might consider simply contributing with your own skillset what the astrophysicists might not have (this is likely what you were asking). In that case, you should join an open source project (they're almost all open source, really). Here are some: http://enzo-project.org/ is numerical code written mainly by cosmologists, but I used it for galactic dynamics research as well with some work (read: a lot of work, but I was a noob). It's written completely in C++. If you know CUDA, you might be very valuable to this project, but that's just one area [that I was interested in]. http://yt-project.org/ is a data analysis software library used with Enzo but also with other simulation software as well. It's written in Python and does some things in parallel. Very cool, I think you might enjoy this one. http://www.mpa-garching.mpg.de/gadget/ is the big project, similar to Enzo. It's an effort headed by Volker Springel, the guy who's paper I linked to in the beginning. I don't know much about this, I just coded some things for http://star-www.dur.ac.uk/~jch/gadgetviewer/. I'm not even entirely sure that it's straightforward to contribute to, but you can check it out and give it a shot. Personally, I recommend the YT project, I think that's most well-suited to your skills. The Enzo project could use some help, but it'll take longer to understand what's going on. I think actually in both cases it will take a while, but since you're interested, I think it should be worth it. Good luck! Maybe after a few years of working on these projects you can tackle the paper I gave above. It would certainly be a very impressive thing to show off to anyone who does any type of computational astrophysics, CFD, nuclear simulations, etc.

There is the discipline of "computational astrophysics".   One place to look is the Virgo Consortium: http://www.virgo.dur.ac.uk/ But note this is not really about abstract computer science, computational theory, etc.  The focus is more on simulations, fluid models, etc.  So the relevant skills would be numeric methods, paralle/distributed computing, etc.

Talk to https://www.zooniverse.org/, they probably have needs that would match your skills.

The problem that you will find is that the astrophysics market is small, and most astrophysicists end up with CS experience.  There are a lot of jobs that you *could* do, but the trouble is that jobs are scarce (and if the jobs were there, I would have never gone off to Wall Street). Something that you can do is to work on some open source astronomy packages.  Download the mageia linux distribution, and install the task-astronomy RPM.  There are probably a dozen things that need to be done.  Most of these tasks are very grunt work, unglamourous things, but there are things that need to be done.  (Look at IRAF, and shudder).

Oh there are many and all of them are amazing! Astrophysics nowadays has reached at such a level that computation and programming is needed at each and every level. I would say that most of the Astrophysicists are programmers at some level.In the last few years, a huge amount of data is being generated from the observational telescopes and missions all over the world and space. Huge computing power, good skills and new optimization techniques are required to get useful results from these data sets. At the same time, simulations on small as well as very large scales are taking place in many branches of Astrophysics. These simulations require good experience in computing and dealing with algorithms. Computer science knowledge definitely comes in handy and sometimes becomes of prime importance to create or optimize these simulations which help us solve mysteries of the Universe. Then there’s also the machine learning part. As time is passing by, this branch is becoming more and more important for the field of Astrophysics. There are many more examples where computer science people contribute immensely to the field of Astrophysics.So overall I’d say, dealing with big data sets, creating and optimizing simulations and, developing Analytical software are the main areas where Computer Science graduates can contribute. It might sound a bit boring, but trust me, once you get into it, you’ll love it! :)