Member of Technical Staff - ML Systems

What you'll do

• Develop infrastructure for distributed post-training and reinforcement-learning experiments, including resource allocation, execution, recovery, and checkpoint evaluation.
• Improve model serving and inference performance, and build scalable data-generation and agent-environment systems.
• Develop research tooling, benchmarks, observability, and automated workflows; help turn methods into dependable training and deployment capabilities.

What they ask for

• Technical depth in ML infrastructure, including experience with training systems, model serving, or distributed data and rollout pipelines.
• Knowledge of how PyTorch, JAX, or similar frameworks work beneath their APIs; ability to profile workloads and diagnose correctness or performance issues.
• Strong distributed-systems fundamentals, including scheduling, communication, storage, and recovery.
• Understanding of post-training execution and ability to evaluate system changes through reproducible measurements.

Nice to have

• Experience with distributed training frameworks, vLLM or SGLang, GPU profiling, container orchestration, or research platforms.
• CUDA, Triton, or compiler and runtime expertise.

Tools

PyTorch, JAX, vLLM, SGLang, CUDA, Triton

AI focus

Build systems supporting model post-training and improvement, including data generation, reinforcement-learning experiments, evaluation, inference, and deployment.