Research Engineer, Cybersecurity RL (Reinforcement Learning)

Helps Anthropic's models get safely better at incident response, security analysis, vulnerability remediation and other defensive security work, as a research engineer working with the company's Security team. The role designs and builds RL environments, runs experiments and evaluations, ships work into production training runs, builds agentic integrations that let AI systems investigate and act on findings, and helps set the team's direction. On-call may be required.

Anthropic wants machine learning experience combined with cybersecurity research and strong software engineering. Applied defensive security work, CTFs or cyber ranges, and familiarity with RL and LLM training help. Based in San Francisco or New York.