Because of my skills in C/C++ and CUDA programming, I believe I would be a fantastic fit for this project. I've had experience working with the CUDA programming model as well as various GPU accelerated libraries that accompany it (mainly cuRand). When I tackle a project that I intend to accelerate with CUDA, I have a tried and true process to achieve that goal. First, I implement the project in a purely serial fashion. Next, I identify the most ideal areas of the code that I can accelerate. In the process of turning them into kernels, I try to parallelize the regions using OpenMP to make sure I can gain better performance via parallelization then I simply turn the OpenMP regions into CUDA kernels. Then, once they're turned into CUDA kernels, I try and see if there is any areas that need optimizing. I hope you give this proposal consideration, and if you do feel free to contact me