Research Engineer, Machine Learning (RL Velocity)

About this role

Anthropic seeks a Research Engineer to build and optimize the reinforcement learning infrastructure that powers the organization's AI research. You'll focus on removing performance bottlenecks, improving training systems reliability, and enabling faster iteration across the research team.

What you'll do

• Develop and enhance RL training infrastructure used by researchers
• Identify and eliminate performance bottlenecks through debugging and profiling
• Collaborate with research and engineering teams to build tools addressing pain points
• Ensure reliability and end-to-end performance of research experiments
• Contribute to architectural decisions for large-scale RL systems

What they're looking for

• Software engineering fundamentals and systems design
• ML infrastructure or distributed systems experience
• Performance optimization and profiling
• ML frameworks (JAX, PyTorch, or similar)
• Full-stack development across algorithms and low-level optimization
• Ability to operate in fast-paced research environments
• Cross-functional collaboration and communication

Benefits

• Competitive annual compensation: $500,000–$850,000 USD
• Remote-friendly with offices in San Francisco, NYC (25% minimum in-office)
• Visa sponsorship support available
• Work on cutting-edge AI safety and reliability research
• Collaborate with top researchers and engineers