Cyber Evaluations Engineer — AI Safety & Abuse Detection

Anthropic; Friendly (MD); Hybrid; USD 300,000 - 405,000; Full time

Benefits offered by this job

Competitive salary

Flexible hours

Office in SF

Equity donation matching

Generous vacation & parental leave

Job summary

Anthropic is seeking Cyber Evaluations Engineers to design and run evaluations that measure cyber-relevant capabilities and safeguard robustness in our models. You will design new evals, run per-release robustness testing, and analyze data on jailbreaks and prompt bypasses to understand where safeguards hold up and where they don’t.

You will also design detection probes, work with the policy team, and help shape a layered, robust abuse-detection architecture while building internal tooling to

Qualifications

• Experience building or running evaluations, benchmarks, or test suites for software or ML systems.
• Hands-on cybersecurity experience (e.g., CTF participation, vulnerability research, exploit development, or security research).
• Proficiency in Python and ability to communicate evaluation results to cross-functional stakeholders.

Responsibilities

• Design and run capability, uplift, and safety evaluations to assess cyber-relevant risk in new models.
• Execute per-release safeguard-robustness testing ahead of major model launches.
• Analyze evaluation results and communicate findings clearly to the team and to stakeholders.
• Design, prototype, and tune detection probes for cyber misuse.
• Collaborate with policy and engineering partners to translate eval findings into safeguard improvements.
• Build and maintain internal tooling used to run and score evaluations.

Skills

Evaluations

Cybersecurity

Python

Stakeholder communication

Education

Bachelor's degree

Job description

Anthropic is seeking Cyber Evaluations Engineers to design and run evaluations that measure cyber-relevant capabilities and safeguard robustness in our models. You will design new evals, run per-release robustness testing, and analyze data on jailbreaks and prompt bypasses to understand where safeguards hold up and where they don’t.

You will also design detection probes, work with the policy team, and help shape a layered, robust abuse-detection architecture while building internal tooling to