Principal Engineer – Scale-Up GPU Networking (HPC / AI)

This role involves designing and implementing GPU-aware networking paths for high-bandwidth, low-latency intra-node communication within HPE's High Performance Computing and AI division. The position focuses on optimizing GPU-to-NIC data movement, integrating with GPU ecosystems like NVIDIA CUDA and AMD ROCm, and enhancing communication stacks such as Libfabric and UCX. The engineer will lead upstream contributions to open-source projects and drive multi-NIC and NUMA performance optimizations for scale-up workflows.