AI Chemical Security Researcher
Posted
Summary
Conducts end-to-end research at the intersection of frontier AI and chemical security: building threat models and benchmarks, red-teaming high-risk chemical tasks, validating safeguards, and turning findings into reports and policy recommendations. Requires chemistry domain expertise plus AI evaluation skills and Chinese language proficiency.
You will conduct end-to-end research on frontier AI and chemical security. You will create threat models and benchmarks, red-team high-risk chemical tasks, validate safeguards, and translate research and policy developments into reports, recommendations, and technical proposals.
Responsibilities
- Develop chemical-security threat models with domain experts
- Define risk thresholds and assessment criteria for large models, chemistry-specific models, AI agents, and automated labs
- Design benchmarks for AI chemical capability and safety
- Assess model performance in chemical knowledge, task execution, and safety risk identification
- Conduct red-teaming evaluations on high-risk chemical tasks
- Identify safety vulnerabilities across models and systems
- Explore and validate risk-mitigation strategies and protective safeguards
- Monitor AI chemical-security research and policy developments
- Convert findings into research reports, policy recommendations, and technical proposals
Requirements
- Master’s degree or above in Organic Chemistry, Medicinal Chemistry, Chemical Engineering, Toxicology, Chemical Biology, or a related field
- AI
- Chemistry
- Technical feasibility evaluation
- Chemical risk evaluation
- Chinese reading, writing, and spoken communication