Software Engineer – Reinforcement Learning Data

About the role

Anthropic is looking for a senior Software Engineer to join its new RL Data team. You will shape the architecture of critical data pipelines, work closely with research groups, and ensure the quality and safety of reinforcement‑learning data that powers Claude.

Key responsibilities

• Own end‑to‑end components of the RL data stack, from architecture to operational maintenance.
• Design and implement data‑collection pipelines, iterate on prompts, evaluations, and grading mechanisms.
• Develop QA frameworks that detect reward‑hacking and guarantee environment integrity.
• Create user‑friendly interfaces for rapid human‑feedback collection.
• Harden execution environments with sandboxing, snapshotting, and comprehensive tool coverage.
• Collaborate with domain experts, operations, security, and compliance teams to roll out systems to new users and vendors.

Required profile

• Proven track record of leading major projects end‑to‑end in fast‑paced, ambiguous settings (e.g., founder, CTO, tech lead, or major open‑source contributor).
• Demonstrated ability to inspire teams, plan workstreams, and proactively resolve blockers.
• Strong software engineering expertise with a modern programming language.
• Experience using AI tools in daily workflows.

Required skills

• Python
• TypeScript
• Docker
• Kubernetes
• Common cloud infrastructure (e.g., AWS, GCP, Azure)