Research Engineer, Cybersecurity RL (Reinforcement Learning)

About this role

Anthropic seeks a Research Engineer to advance reinforcement learning for cybersecurity applications, combining expertise in secure coding and vulnerability remediation with AI safety. The role bridges research innovation and production engineering within the Horizons team, requiring both novel RL approach development and hands-on implementation.

What you'll do

• Design and implement reinforcement learning environments for cybersecurity tasks
• Conduct experiments and evaluate RL-based approaches for secure coding and vulnerability remediation
• Develop novel RL techniques tailored to defensive cybersecurity domains
• Integrate research outcomes into production AI model training pipelines
• Collaborate with researchers, engineers, and cybersecurity specialists across teams
• Balance exploratory research with practical software engineering implementation

What they're looking for

• Cybersecurity research and domain knowledge
• Reinforcement learning techniques and environments
• Machine learning and deep learning
• Strong software engineering and implementation
• Python or similar programming languages
• LLM training methodologies
• Security engineering and defensive workflows
• Research design and experimentation

Benefits

• Annual salary range: $300,000–$405,000 USD
• Visa sponsorship available
• Hybrid work policy (minimum 25% office time)
• Work on cutting-edge AI safety and beneficial AI systems
• Collaboration with world-class researchers and engineers