Principal Offensive Security Agent Engineer

OpenAI; San Francisco (CA); On-site; USD 347,000 - 490,000; Full time

Job summary

OpenAI seeks a Staff - Principal level offensive security domain expert to build production-grade agents that identify, validate, and drive remediation of vulnerabilities across OpenAI's infrastructure and applications. You will own the architecture, evaluation strategy, and integration of agent systems at scale.

As automation expands, you will design specialized agents that understand OpenAI's environment, testing cloud, Kubernetes, web apps, endpoints, and external surfaces with guardrails and

Qualifications

• Substantial hands-on offensive security experience with strong judgment on viable vulnerabilities and attack paths.
• Domain expertise in cloud, Kubernetes, and modern web applications is highly valued.
• Experience assessing complex, customized environments beyond standard scanners.
• Ability to translate expert workflows into reliable software and agent systems.
• Proven track record building production-quality software and agent-driven tooling.

Responsibilities

• Serve as the technical owner of OpenAI's offensive security agents, defining architecture and strategy.
• Design and build a portfolio of agents that test infrastructure and applications from multiple perspectives.
• Translate workflows into tools, harnesses, policies, and internal knowledge bases.
• Develop capabilities for testing cloud, Kubernetes, web apps, and external attack surfaces.
• Create end-to-end vulnerability-management loops including remediation verification.
• Design human-in-the-loop controls to authorize or adjust potentially dangerous actions.
• Create feedback mechanisms for learning from domain experts and improving agent behavior.
• Develop rigorous evaluations measuring security outcomes and agent capability improvements.
• Build production-quality infrastructure that runs continuously with observability and safe production operation.
• Investigate failures in agent reasoning and improve tools, context, and guardrails.
• Collaborate with security and engineering teams to ensure findings are actionable.
• Help define the future of offensive security at OpenAI with scalable agent-assisted testing.

Skills

Offensive security

Cloud security

Kubernetes security

Web application security

Agent systems

Tools

Kubernetes

Cloud security tooling

Threat modeling tools

Job description

OpenAI seeks a Staff - Principal level offensive security domain expert to build production-grade agents that identify, validate, and drive remediation of vulnerabilities across OpenAI's infrastructure and applications. You will own the architecture, evaluation strategy, and integration of agent systems at scale.

As automation expands, you will design specialized agents that understand OpenAI's environment, testing cloud, Kubernetes, web apps, endpoints, and external surfaces with guardrails and