Research Engineer, RL Engineering

Sourced from aihiringboard.com

Anthropic is hiring a Research Engineer for its RL Engineering team, which builds and owns the reinforcement learning training system behind Claude. You will work across the full stack—orchestration, environments, training, inference, and evaluation—to improve how RL training behaves at scale for both production and research runs. The role sits at the center of RL at Anthropic, collaborating closely with research teams across the company.

• Build, own, and improve the core RL training system serving production and research runs
• Work across the stack: orchestration, environments, training, inference, and evaluation
• Study how RL training behaves at scale and contribute to research that improves it
• Implement new training methods as stable, fast, well-tested code
• Improve speed and efficiency of RL training and evaluation through profiling, optimization, and benchmarking
• Make the system easier for researchers via clean abstractions, clear APIs, and automated testing
• Debug hard problems across the stack, including distributed systems failures at scale
• Communicate results clearly in writing and discussion

• Proficiency in Python and experience debugging and improving large ML codebases