Research Engineer, ML Infrastructure

Highlights

• Own distributed training systems running across thousands of GPUs.
• Scale hundreds of thousands of concurrent coding agent rollouts.
• Build tooling that turns raw compute into faster research iterations.

Ideal candidate

Deep systems thinker who thrives beside researchers and owns infrastructure end to end. You care about speed, reliability, and turning massive compute into real research velocity at frontier scale.

Requirements

• Deep experience building and operating distributed training systems for large models, owning infrastructure end to end
• Strong systems engineering fundamentals across distributed systems, networking, storage and full-stack performance
• Proficiency in Python and C++ with PyTorch or equivalent frameworks at a systems level
• Hands-on GPU performance profiling, memory optimization and compute efficiency to fix underperforming runs
• Experience implementing parallelism strategies including data, tensor, pipeline and sequence parallelism
• Track record building tooling and abstractions that accelerate research workflows
• Strong debugging instincts across complex distributed systems with non-deterministic failures
• Enough ML knowledge to engage substantively with researchers on architectures and infra needs