Threat Modeller

Highlights

• Build threat models that calibrate dual-use risks for Gemini releases.
• Partner across teams to evaluate CBRNe risks in high-capability models.
• Monitor emerging dual-use methods and frontier AI competitor landscape.

What they offer

• 15% bonus target
• equity
• benefits

Ideal candidate

Operational doer with science background who thrives absorbing complex data under pressure and translating high-consequence CBRNe risks into clear action to keep frontier AI safe.

Requirements

• PhD in Science, Engineering, Data Science or equivalent practical experience
• 3 years of experience designing qualitative and quantitative models
• 1 year working in national laboratory, research organization, government defense, military intelligence or specialized research organization
• Experience executing independent projects and synthesizing large datasets under tight deadlines
• Experience red-teaming and evaluating LLMs to identify vulnerabilities and misuse
• Strong knowledge of dual-risk CBRNE components and laboratory/defense paradigms
• Understanding of model evaluation methodologies for frontier AI risks including FSF