H-100 GPU Waveform test

Our law firm has a client that has developed a new software company that we need to test independently. The company is an early-stage tech company with an existing codebase written by our CTO. It needs to be built, run, and validated on an NVIDIA H100. We’re looking for an experienced GPU engineer to set up the environment, get the code running end to end, and report back on correctness and performance. Scope of work Provision or use an H100 instance (80GB, SXM or PCIe). This can be your own cloud account, such as Lambda, RunPod, CoreWeave, AWS p5, or GCP A3, or one we provide. Set up the environment: CUDA toolkit, drivers, Python environment, and any framework dependencies (e.g., PyTorch, Triton, or custom CUDA kernels). Build the code and resolve any dependency, compilation, or version-compatibility issues. Run the provided test suite and sample workloads. Profile performance using Nsight Systems/Compute or equivalent. Report on GPU utilization, memory usage, throughput, and any bottlenecks. Document every bug, crash, or numerical discrepancy you find, with steps to reproduce. Deliver a written report plus a reproducible setup (Dockerfile or setup script).