Hands-on research internship in system software and I/O architecture for scalable multi-GPU platforms. Contribute to OS/runtime/compiler co-optimization, CUDA performance, profiling, and energy-efficient, high-throughput storage access for accelerated workloads.