Senior Staff Machine Learning Performance Engineer
Core
Own training speed for foundation-model training across thousands of accelerators, setting technical direction for performance and building agentic workflows to automate optimizations.
Role type
Senior Staff Machine Learning Performance Engineer
Builds
Modular performance components, observability tooling, and agentic automation systems for ML training stacks.
Domain
Autonomous driving, robotics, and large-scale AI training
Deliverable
production ML models
Required skills
GPU training/inference optimization, computer architecture, high-performance computing, C++, Python, data structures, algorithms, multi-node scaling, production debugging
Preferred skills
CUDA/Triton kernel tuning, PyTorch internals, profiling/tracing infrastructure (CUPTI, Perfetto), LLM agents for automation, autonomous driving/robotics domain expertise
Technologies
CUDA, Triton, PyTorch, CUPTI, Perfetto
Responsibilities
Optimize step time, scaling efficiency, and stability for foundation-model training; collaborate with modeling and infrastructure teams to resolve bottlenecks; build reusable performance modules; extend observability tooling; develop agents to automate job reproduction and regression detection.
Seniority
Senior Staff, hands-on IC with technical leadership