I'd prefer someone working with graphics, since until it's done I don't have any pretty pictures to show people . The graphics library is 2D vector based, so the idea is that the main simulation display, graphics, etc. should all allow you to freely navigate around and zoom in/out without things looking horrible. Even just going through and tightening up the API a bit would be useful, as would writing some tests for parts of the graphics not yet tested.
Another idea: you could do a module to draw graphs. I've tried googling around and I haven't found anything terribly compelling for it. Basically it would just take a set of data points (Scalar x,y pairs (vectors, essentially)), maybe do some statistics on the data (best fit curves, etc.), and then draw it to a window handle using the graphics library. Maybe also some graph visualization where you create a family tree for a bot and then the graphics library would use a energy minimization function to arrange the tree efficiently. For the math, you could build on top of the math library I've been working on (Azimuth) easily enough, and that would let me go in later and do fun mathy stuff with the data. On the graphics side, someone could go in later and add post processing effects like a glow shader (everything looks better
glowing).
The advantage of a project like this is that it's fairly independent from the rest of the engine, so you should be able to work without figuring out the whole code base at once and relying on me too much to fix issues you run in to. You can also provide feedback on the graphics API as it stands, which is pretty horrible usability-wise.