Research Program Manager, Frontier Assurance
Remote • San Francisco • FullTime
Posted 1d ago
Remote Work Policy
Fully remote
Employment Type
FullTime
Categories
AI Research Engineer
About the job
OpenAI is seeking a Research Program Manager for its Frontier Assurance team. This role is pivotal in building programs that integrate independent expertise into frontier AI safety decisions and communicate the evidence behind these decisions to the public. You will lead external research partnerships, third-party assessments, and coordinate public safety documentation, while developing novel approaches to independent scrutiny and transparency. This position requires close collaboration with research, engineering, product, policy, and communications teams to ensure external findings influence concrete decisions and that public communications accurately reflect the evidence, limitations, and uncertainties.
Responsibilities
- Design and manage third-party assessment programs for frontier models and safeguards, including evaluations, adversarial testing, and embedded auditing.
- Build and manage strategic research partnerships with external evaluators, academic labs, and independent experts.
- Translate external findings into actionable recommendations for safety decision-makers, influencing safeguards, deployment, product, and policy.
- Partner with internal teams to communicate the evidence behind OpenAI’s safety approach, including testing methodologies, learnings, and impact on decisions.
- Lead public transparency initiatives, such as system cards and summaries of third-party assessments.
- Establish channels for external researchers and civil society to provide feedback and inform future assessments and transparency efforts.
- Own program goals, milestones, dependencies, and risks, ensuring stakeholder alignment and resolving execution obstacles.
Requirements
- Understanding of AI evaluations and measurement, with the ability to engage technical teams on evaluation design and results.
- Ability to interpret evaluation findings, including methodological limitations, uncertainty, and disagreement.
- Experience building research partnerships and managing external stakeholders, particularly academic researchers and independent experts.
- Skill in synthesizing technical and social science research into clear executive summaries and accessible public explanations.
- Experience delivering technical publications or public reports with multiple contributors.
- Experience working cross-functionally across product, research, and engineering teams.
- Ability to coordinate complex programs involving policy and communications partners.
- Understanding of and interest in frontier AI safety and policy.
Benefits
- Hybrid work model (3 days in office per week)
- Relocation assistance for new employees