Senior HPC and AI Network Software Architect

NVIDIA, a leader in graphics, PC gaming and accelerated computing, is expanding its AI efforts and seeks a Senior HPC and AI Network Software Architect to shape the next generation of scalable AI infrastructure. The position focuses on distributed training, real‑time inference and efficient communication across large‑scale systems that power the world’s fastest AI workloads. In this role you will design and evolve software architectures that support cluster‑wide AI training and inference, emphasizing throughput, latency, resiliency and memory efficiency. You will create and assess new communication and runtime features in libraries such as NCCL, UCX and UCC, and work closely with framework teams like TensorFlow, PyTorch and JAX to integrate and improve end‑to‑end performance. Collaboration with hardware groups will involve GPUs, DPUs and interconnects to accelerate data movement, while you drive innovations in runtime systems, communication stacks and AI‑specific protocols, turning concepts into robust production code. Candidates must hold a Ph.D. or comparable industry experience in computer science, computer engineering or a related discipline, and bring at least five years of systems programming, parallel or distributed computing, high‑performance networking or large‑scale data movement experience. Proficiency in C++, Python and preferably CUDA or other GPU programming models is required, along with a proven record of delivering performance‑critical software. Hands‑on familiarity with AI frameworks and a deep understanding of how communication libraries enable large‑scale training and inference are essential, as are strong collaboration skills in a multinational, interdisciplinary environment. Additional strengths include expert knowledge of NCCL, UCX, UCC or similar libraries, experience with RDMA, collective communications and accelerator‑aware networking, and a background in hardware‑software co‑design for distributed AI. Familiarity with large‑model deployment techniques such as sharding or pipeline parallelism is also valued. NVIDIA offers a supportive, learning‑focused culture, competitive compensation and a comprehensive benefits package for you and your family.