Software Engineer, Guardian
Core
Design and build production systems that evaluate AI agent behavior, identify risks, and intervene to prevent harm while preserving user intent.
Role type
Senior IC software engineer (AI safety/oversight systems)
Builds
Auto-review platform and broader Guardian oversight systems for OpenAI AI agents
Domain
AI safety, agentic systems, backend infrastructure
Deliverable
production ML models
Required skills
backend service design, API development, infrastructure design, real-time system evaluation, adversarial testing, privacy protection, product-engineering judgment, system reliability, failure investigation, technical direction
Preferred skills
experience building with AI agents or LLMs, translating research into scalable systems, working with alignment researchers, navigating ambiguity
Technologies
(none explicitly listed)
Responsibilities
Design, build, and operate core components of Auto-review and the Guardian platform; Build production systems to evaluate agent actions in real time; Improve oversight systems' understanding of user intent and context; Partner with researchers to turn prototypes into scalable capabilities; Develop evaluations and adversarial tests to monitor safety performance; Own projects end-to-end from exploration to production deployment
Seniority
Senior, hands-on IC
