Cybersecurity Capabilities Researcher for AI Security Products

Anthropic Limited; San Francisco (CA); Hybride; USD 405 000 - 485 000; Plein temps

Résumé du poste

Anthropic is seeking a Capabilities Researcher to join the Claude Security team in San Francisco. You’ll identify security capabilities, measure performance, and prototype solutions for non-experts, shaping what the model can do in real-world security contexts.

You’ll collaborate with engineers and researchers to design scaffolding, tooling, and defaults that translate frontier research into practical capabilities, with a strong emphasis on rigorous evaluation and responsible AI safety.

Qualifications

• Bachelor’s degree or equivalent combination of education, training, and/or experience.
• 7+ years of experience in security research, security engineering, or a closely related field.
• Ability to design rigorous evaluations and interpret results honestly; communicate technical findings clearly.

Responsabilités

• Prototype rapidly to define the AI frontier for cybersecurity work.
• Design evaluations that measure model performance on security tasks.
• Build datasets, harnesses, and scoring those evaluations depend on.
• Engage with the cybersecurity community to define impact opportunities.
• Collaborate with engineers and researchers to operationalize capabilities into products.
• Track how model capabilities for security are changing and inform future work.
• Share findings that inform product direction and prioritize work with leadership.

Connaissances

Security research

Security engineering

Reverse engineering

Exploit development

Malware analysis

Incident response

Red teaming

AI security

Communication skills

Formation

Bachelor’s degree

Description du poste

Anthropic is seeking a Capabilities Researcher to join the Claude Security team in San Francisco. You’ll identify security capabilities, measure performance, and prototype solutions for non-experts, shaping what the model can do in real-world security contexts.

You’ll collaborate with engineers and researchers to design scaffolding, tooling, and defaults that translate frontier research into practical capabilities, with a strong emphasis on rigorous evaluation and responsible AI safety.