Research Engineer - Distributed Training

Highlights

• Build distributed training behind pre-training and large-scale RL workloads.
• Shape architecture of async rollout and post-training RL stack.
• Contribute to open-source libraries for frontier-scale model training.

What they offer

• Cash compensation range of $150-350k plus equity incentives
• Flexible work arrangements with remote or in-person option in San Francisco
• Visa sponsorship and relocation assistance for international candidates
• Quarterly team off-sites, hackathons, conferences and learning opportunities

Ideal candidate

Systems-minded engineer who loves squeezing performance from large-scale training and thrives on ambiguous, high-ownership problems. Excited by open frontier AI and collaborating closely with researchers to build real infrastructure.

Requirements

• Strong systems engineering experience in AI/ML infrastructure for large-scale training or inference
• Deep familiarity with PyTorch and distributed training frameworks
• Experience optimizing training performance across kernels, memory, communication and parallelization
• Hands-on experience with data parallelism, tensor parallelism and pipeline parallelism
• Strong understanding of GPU architecture, profiling and performance debugging
• Ability to identify bottlenecks across the stack and drive improvements from first principles
• Comfort working in fast-moving environment with ambiguous problems and high ownership