Software Engineer, Cybersecurity Benchmarks
Remote
Core
Build and maintain cybersecurity benchmarks for frontier AI models, secure evaluation environments against AI escape, and shape evaluation strategy using threat modelling.
Role type
Senior IC Research Engineer (AI Security Benchmarks)
Builds
Public AI cybersecurity benchmarks and secure evaluation sandboxes
Domain
AI Safety + Cybersecurity
Deliverable
production ML models
Required skills
Software engineering, red teaming, sandboxing, threat modelling, AI coding harnesses, Inspect library
Preferred skills
Experience with frontier AI models, infrastructure scaling
Technologies
Inspect library, Python, cloud infrastructure
Responsibilities
Design and maintain cybersecurity benchmarks, harden sandboxes to prevent AI escape, prototype new benchmark ideas, integrate with AI providers, collaborate with researchers on evaluation data
Seniority
Senior, hands-on IC
