Offensive Security Agent Engineer

OpenAI is an AI research and deployment company dedicated to ensuring that artificial general intelligence benefits all of humanity. The Staff–Principal Offensive Security Agent Engineer will own the architecture and development of production-grade agents that continuously identify, validate, and coordinate remediation of vulnerabilities across infrastructure and applications. The role also involves building evaluations, safety controls, feedback mechanisms, and reliable systems for autonomous security testing.

Responsibilities:

• Serve as the technical owner of OpenAI’s offensive security agents, establishing its architecture, technical direction, operating model, and evaluation strategy
• Design and build a portfolio of specialized agents that continuously test OpenAI’s infrastructure and applications from a variety of authenticated and unauthenticated perspectives
• Translate expert offensive security workflows and intuition into tools, skills, harnesses, policies, and internal knowledge bases
• Build agents that deeply understand OpenAI’s environment by integrating internal context
• Develop capabilities for testing cloud and Kubernetes environments, modern web applications, external attack surface, endpoints, and other high-value systems
• Build complete vulnerability-management loops that move beyond discovery to impact validation, ownership identification, prioritization, remediation support, progress tracking, and fix verification
• Design human-in-the-loop systems that allow offensive security engineers to approve or reject potentially dangerous actions, provide missing context, redirect investigations, and steer agents away from unproductive paths
• Create feedback mechanisms that allow agents to learn from the decisions, corrections, and domain expertise of experienced offensive security practitioners
• Develop rigorous evaluations that measure meaningful security outcomes and improvements in agent capability over time
• Build production-quality infrastructure that allows the system to run continuously, recover from failures, remain observable and debuggable, and operate safely against production systems
• Investigate failures in agent reasoning and behavior, identify where models are capable or unreliable, and improve the surrounding tools, context, workflows, and guardrails accordingly
• Partner closely with offensive security, infrastructure security, product security, codex security, and engineering teams to ensure findings are high signal, understandable, and actionable
• Help define the future of offensive security at OpenAI, with the goal of enabling agents to perform most repeatable security testing while human experts focus on automation and high leverage agent-assisted manual review

Requirements:

• • You have substantial hands-on offensive security experience and strong judgment about which vulnerabilities and attack paths are worth pursuing
• • You have extensive domain expertise in areas such as cloud security, Kubernetes and container security, web application security, source-code review, Linux security, macOS security, or external attack-surface testing. Expertise in cloud, Kubernetes, and modern web applications is especially valuable
• • You have experience assessing complex, highly customized environments rather than relying primarily on standardized scanners, checklists, or known-vulnerability detection
• • You can take an ambiguous offensive security problem, decompose it into a reliable system, and encode the reasoning and workflows of an experienced operator into software
• • You have built production quality software
• • You have built or meaningfully extended agent systems that use models, tools, structured context, memory, orchestration, and feedback loops to perform complex work
• • You understand that an impressive agent demonstration is very different from a dependable production system, and you care deeply about evaluations, observability, failure recovery, safety, maintainability, and regression resistance
• • You have strong intuitions about where current models are capable, where they are unreliable, and how tools, context, scaffolding, and human feedback can expand their useful operating range
• • You are excited about working closely with frontier models, curious about their emerging capabilities, and constantly look for ways to use them to improve your own workflows
• • You are energized by the opportunity to serve as a technical owner of an ambitious new system, make foundational architectural decisions, and help grow a team around it
• • Background or expertise in AI or data science
• • Prior experience working in tech startups or fast-paced technology environments
• • Experience in related disciplines such as Software Engineering, Product Security, Application Security, Detection Engineering, Site Reliability Engineering, Security Engineering, or IT Infrastructure