Researcher, Agent Safety, Oversight and System Mitigations

The role

This researcher will focus on oversight and system-level mitigations to enable capable agents to operate safely and autonomously. You will design, build, and evaluate system-level controls and red-team agentic systems.

What you will do

• Design, build, and evaluate system-level controls for agent actions.
• Plan how controls fit in a broader system including sandboxing and permission boundaries.
• Work closely with a Codex harness engineering team to productionize AI controls.
• Red-team end-to-end agentic systems to measure control effectiveness.
• Improve the safety–productivity tradeoff by measuring and reducing harmful actions, blocks, burden, and latency.

Required qualifications

• Safety and security minded researcher or engineer.
• Ability to reason about security boundaries and agent behavior.
• Ability to build and test practical mitigations.
• Strong systems or security instincts.
• Ability to reason about isolation boundaries, permissions, attack surfaces, and failure modes in complex systems.
• Ability to turn ambiguous safety questions into concrete threat models, reproducible experiments, and practical mitigations.
• Ability to build robust experimental infrastructure and design evaluations.
• Deep interest in frontier AI alignment, safety and control.

Nice to have

• Background in AI control or security

Benefits

• Relocation assistance

Before you apply

• Background checks for applicants will be administered in accordance with applicable law.

About the company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity.