async Python Jobs

Anthropic

Reinforcement learning is crucial for training AI models to reason, write code, and act autonomously. At scale, this process involves a demanding distributed system where training, sampling, and environment execution run concurrently across a large fleet of accelerators and hosts. This system must continuously exchange data and maintain progress despite hardware failures, shifting loads, and evolving research. The efficiency of this system directly impacts learning outcomes and the speed of research iteration. As a Research Engineer on the Distributed Systems team within RL Engineering, you will focus on the most critical bottlenecks within this system. This could involve areas like scheduling, data movement, sandboxed environment execution, storage, networking, fault tolerance, autoscaling, or observability. The role seeks generalist engineers capable of reasoning from first principles across different system layers and prioritizing impactful problems over familiar ones. The ideal candidate will have experience building and operating large distributed systems, a commitment to detail, and a desire to apply their expertise to a complex, heterogeneous, and rapidly changing workload.

$500k - $850k