Researcher, Cybersecurity Products
$320k - $405k • San Francisco, CA
Posted 11h ago
Job Location
San Francisco, CA
Tech Stack
Remote Work Policy
On-site
Categories
Applied AI Engineer
About the job
Anthropic is seeking a Capabilities Researcher to join the Claude Security team. This role involves identifying, measuring, and refining security capabilities in advanced AI models to make them accessible and useful for non-security experts. You will investigate which security capabilities are reliable, how they perform under realistic conditions and adversarial involvement, and their limitations. Your work will involve rapid prototyping and rigorous evaluation to answer these questions and will directly influence the team's product development decisions. You will also collaborate with engineers to design the necessary infrastructure, tools, and defaults to translate model capabilities into practical applications for customers, remaining involved as these capabilities evolve into products. This is a research-focused position within a product team, offering significant autonomy to explore promising AI applications for security.
Responsibilities
- Prototype rapidly to define the AI frontier for cybersecurity work.
- Design evaluations to measure model performance on tasks relevant to security teams.
- Build datasets, harnesses, and scoring mechanisms for evaluations.
- Engage with the cybersecurity community to identify areas for AI impact.
- Collaborate with engineers and researchers to operationalize promising capabilities for customer use.
- Track changes in model security capabilities and their implications for future development.
- Share findings to inform product direction and partner with product leadership on priorities.
Requirements
- Deep expertise in one or more security domains (e.g., vulnerability research, exploit development, reverse engineering, malware analysis, incident response, offensive security).
- Experience building AI-powered tools or capabilities for security work.
- Ability to quickly prototype ideas and abandon unpromising ones.
- Proficiency in designing rigorous evaluations and interpreting results.
- Clear written and verbal communication skills for technical findings.
- 7+ years of experience in security research, security engineering, or a related field.
- Published research, CTF results, CVEs, or open-source security tooling (preferred).
- Experience with model evaluation, benchmarking, or red teaming (preferred).
- Experience building agentic applications (preferred).
- Familiarity with AI safety considerations in security contexts (preferred).