Senior Software Engineer, AI Inference Systems

This role is about building high-performance AI inference systems using NVIDIA GPUs. You will work on vLLM, an open-source inference framework, optimizing it for the newest models and hardware. Daily work includes writing custom GPU kernels (hand-tuned and compiler-generated), developing compiler infrastructure, and defining benchmarking methodologies like MLPerf. You'll also architect scheduling and orchestration for large-scale multi-GPU, multi-node deployments across clouds. The position combines deep systems engineering with research, pushing the frontier of accelerated computing for AI.