Model Policy Manager, Agentic Safety

Highlights

• Shape how OpenAI understands real-world risks from misaligned agentic models.
• Investigate failures across extended trajectories and turn them into safeguards.
• Build monitoring to detect regressions and emerging risks after deployment.

What they offer

• Relocation support
• Equity

Ideal candidate

Adversarial-minded safety researcher passionate about AI alignment who thrives turning ambiguous agentic failures into concrete, empirically grounded safeguards in fast-moving teams.

Requirements

• Strong background in AI agent safety, privacy, security, cybersecurity, or adjacent fields with adversarial mindset
• Demonstrated interest in AI alignment and understanding of technical drivers of misaligned behavior
• Technical fluency to work directly with evaluation and training data and identify patterns and limitations
• Hands-on with model data and evaluation results, inspecting failures and assessing data quality
• Uses empirical evidence to develop and refine safety policies and safeguards
• Can translate ambiguous alignment risks into precise behavioral expectations and evaluation criteria
• Works effectively across research, engineering, security, product, and policy
• Communicates clearly about complex and uncertain technical risks