Staff Software Engineer – Code RL

About the role

Anthropic is building safe, interpretable AI systems and needs a Staff Software Engineer to lead the Code RL effort behind Claude's coding capabilities. You will work closely with research teams, designing and scaling agentic coding environments, and ensuring the reliability of production reinforcement‑learning pipelines.

Key responsibilities

• Design and implement widely‑used APIs, frameworks, and abstractions for researchers and engineers.
• Embed with research teams on a rotational basis, understand their needs, build supporting systems, and hand over ownership.
• Improve reliability and structure of research codebases without slowing down ongoing research.
• Anticipate silent failure modes and mitigate them through type safety, invariants, targeted testing, and refactoring.
• Contribute to production RL system health, including monitoring, regression detection, and triage tooling.
• Help define engineering standards, review practices, and mentor others in adopting them.

Required profile

• Deep expertise in Python, including static typing, safe async, and concurrency patterns.
• Proven experience designing intuitive, safe APIs or frameworks used by multiple teams.
• Strong intuition for how complex systems fail, especially silently, and ability to prevent such failures.
• Comfort working directly in messy research code and iteratively improving core abstractions.
• Ability to mentor engineers and researchers on engineering best practices.

Required skills

• Python
• Static typing
• Async programming
• Concurrency patterns
• API design
• Framework design
• Monitoring
• Testing
• Regression detection
• Triage tooling