Requirements

• Experience building and shipping complex production software with end-to-end ownership.
• Familiarity with AI agents or large language models in a core product, startup, or independent project.
• Strong software engineering fundamentals, including design of reliable backend services, APIs, and infrastructure.
• Sound product-engineering judgment balancing safety, usability, speed, and reliability.
• Ability to work closely with researchers to translate ideas into practical systems using experiments and evaluations.
• Comfortable with ambiguity, proactive, and able to communicate clearly across technical disciplines.
• Deep commitment to building AI systems that are useful, trustworthy, and safe.

What You'll Be Doing

• Design, build, and operate core components of Auto-review and the Guardian platform balancing safety, utility, latency, and reliability.
• Develop production systems that evaluate agent actions in real time, detect harmful behavior, and support safe recovery or escalation.
• Enhance oversight systems to understand user intent, environmental context, permissions, and personalization while protecting privacy.
• Collaborate with alignment researchers to scale prototypes, evaluations, and monitoring into product capabilities.
• Work across Product, Security, Infrastructure, and engineering teams to integrate Guardian into OpenAI agent experiences.
• Create evaluations, adversarial tests, and monitoring to analyze safety performance, false positives, latency, and user impact.
• Take end-to-end ownership of projects, contribute to technical design and reviews, and investigate failures to improve systems.