Principal Architect, AI Networking

What you'll do

• Set the technical vision for distributed AI communication and lead research and prototyping of networking solutions.
• Drive hardware-software co-optimization and integrate networking capabilities into AI serving stacks.
• Publish research, represent NVIDIA in industry forums and standards bodies, and mentor senior engineers.

What they ask for

• 15+ years in systems software and/or networking, with deep expertise in high-performance networking, communication libraries, and GPU-accelerated systems; track record delivering complex initiatives from research to production.
• Deep understanding of computer architecture, memory hierarchies, DMA engines, and OS-level networking.
• Understanding of ML systems concepts such as transformer architectures, KV cache mechanics, model parallelism, or distributed training and inference.
• Proficiency in C, C++, Rust, and Python.

Nice to have

• Knowledge of ML inference frameworks and their communication requirements; CUDA programming and NVIDIA GPU architecture expertise.
• Experience influencing product strategy and technical roadmaps at a senior level; major open-source contributions.

Pay

272,000 USD - 431,250 USD (base salary)

Tools

RDMA, NVLink, GPUDirect, InfiniBand, RoCE, NCCL, UCX, MPI

AI focus

Researches and architects communication systems for large-scale AI workloads, including GPU-to-GPU and GPU-to-storage data movement, AI inference and serving, model parallelism, and distributed training and inference.