Researcher, Agent Safety, Training and Evaluations
OpenAI · San Francisco
About this role
**Researcher, Agent Safety, Training and Evaluations** **About the Team** The Agent Safety team ensures increasingly capable AI agents act safely, exercise sound judgment, and remain aligned with user intent. Work spans three areas: • **Training**: Create methods, environments and data that teach agents to make better decisions in consequential situations • **Measurements**: Build evaluations and metrics that identify emerging risks • **Oversight**: Develop oversight mechanisms that reduce harmful actions while preserving useful autonomy **About the Role** We're seeking strong executors with excellent judgment, comfort with ambiguity, and understanding of frontier model research. Prior safety experience not required. **Location**: San Francisco, CA (Hybrid - 3 days/week in office, relocation assistance available) **Key Responsibilities** • Train and evaluate frontier models to reduce harmful or misaligned agent actions • Mine incidents and build scalable measurement and evaluation systems • Collaborate with post-training, capabilities, oversight, and pre-training teams to ship mitigations **Ideal Candidate** • Demonstrated strength in research engineering, ML engineering, quantitative research, or applied model research • Strong technical execution in experimentation, data, evaluation, and/or infrastructure • Motivated by agent safety and eager to work on urgent, practical problems **About OpenAI** An AI research and deployment company dedicated to ensuring general-purpose AI benefits all of humanity. We are an equal opportunity employer. *For accommodations or compliance concerns, see full posting.*
Listing freshness
CronJobs last confirmed this listing 19h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.