Threat Modeller, DeepMind
New York, NY, USA +1 · Full-time · Posted Oct 12, 2026
- Location
- New York, NY, USA
London, UK - Experience
- 3 years designing qualitative and quantitative models; 1 year in one of the listed research, government, defense, or intelligence organizations.
- Education
- PhD in Science, Engineering, Data Science, or equivalent practical experience.
About the job
Build and maintain threat models and evaluations to assess and mitigate CBRNe-related dual-use risks in advanced AI models, informing safety reviews and deployment decisions.
What you'll do
- Refine the threat-modeling framework and integrate it into FSF reviews to support deployment decisions.
- Partner with evaluation and mitigation teams to identify CBRNe risks and assess whether safeguards are adequate.
- Engage external stakeholders and monitor emerging dual-use methods and the frontier AI landscape.
What they ask for
- PhD in Science, Engineering, Data Science, or equivalent practical experience.
- 3 years of experience designing qualitative and quantitative models.
- 1 year of experience in a national laboratory, research organization, government defense organization, military intelligence unit, or specialized research organization.
Nice to have
- Experience red-teaming or evaluating LLMs to identify vulnerabilities and potential misuse.
- Strong knowledge of dual-risk CBRNE components and laboratory or defense paradigms.
Pay
$174000 - $252000 (USD) base salary
Tools
Gemini, LLMs, FSF
AI focus
Threat modeling and evaluating CBRNe-related dual-use risks in advanced AI models, and assessing safeguards for safe deployment.
Model-written summary · the original posting is authoritative