SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Anthropic is seeking a Research Engineer to join its Security team, blending research and engineering expertise to advance AI safety in cybersecurity domains. The role focuses on developing new approaches to improve model capabilities in secure coding, vulnerability remediation, and defensive cybersecurity tasks.
Key responsibilities include designing and implementing reinforcement learning (RL) environments and evaluations that measure and improve model performance in security-related tasks. You will run experiments to develop and test new training approaches, then deliver successful results into production training runs. Collaboration is central: you'll partner with researchers, engineers, and security specialists across Anthropic to identify where AI can strengthen defensive workflows, and work with external cybersecurity experts to ground environments and evaluations in realistic defensive work.
You'll contribute to strategic direction decisions about what to build, what partnerships to pursue, and where to invest team resources. The role includes participation in an on-call rotation as needed.
Minimum qualifications include demonstrated experience in cybersecurity research or applied security work, combined with machine learning experience. You must be able to balance research exploration with engineering implementation and care deeply about AI safety and beneficial systems development.
Preferred qualifications include professional experience in security engineering, fuzzing, detection and response, or other applied defensive work; experience building or participating in CTF competitions and cyber ranges; academic research in cybersecurity; familiarity with RL techniques and environments; and knowledge of LLM training methodologies.
Anthropologic requires a Bachelor's degree or equivalent combination of education, training, and professional experience in a field relevant to the role. The position operates on a hybrid model with expectations that staff be in office at least 25% of the time, though some roles may require more. Visa sponsorship is available.
