Member of Technical Staff - Inference

Role overview Build the systems that make large language models practical to serve and evaluate at scale. This role spans cloud inference, performance optimization, and reinforcement learning infrastructure, connecting model-serving systems with the training workflows that use them.

Responsibilities - Develop multi-tenant model-serving infrastructure across cloud GPU fleets. - Design placement and scheduling for different accelerator types. - Improve reliability and resilience in distributed inference services. - Optimize inference systems and integrate them into reinforcement learning workflows. - Help teams evaluate and serve models produced by large-scale training.

Requirements - Experience building or operating large-scale model-serving or inference systems. - Strong understanding of GPU computing and accelerator-aware scheduling. - Ability to work across cloud infrastructure, inference performance, and machine learning training systems. - Practical skills in designing reliable distributed services. - Interest in contributing to open development and sharing technical work where appropriate.

Benefits and work setup - Full-time role, based remotely or in San Francisco.