Lead Offensive Security Agent Engineer

OpenAI; New York (NY); On-site; USD 260,000 - 360,000; Full time

Job summary

OpenAI in New York seeks a Staff- Principal offensive security domain expert to architect and own production-grade security agents that continuously identify vulnerabilities, validate exploits, drive remediation, and evolve with guardrails.

You will build a portfolio of agents, integrate context, test cloud, Kubernetes, web apps, endpoints, and attack surfaces, and collaborate with security and engineering teams to deliver high-signal, actionable findings at scale.

Qualifications

• Hands-on offensive security experience with strong judgment on attack paths.
• Domain expertise in cloud security, Kubernetes, web apps, and external attack surface testing.
• Ability to translate expert workflows into reliable software and guardrails.

Responsibilities

• Serve as technical owner of OpenAI’s offensive security agents and define architecture and direction.
• Design and build agents testing OpenAI's infrastructure and applications from multiple perspectives.
• Translate offensive security workflows into tools, policies, and knowledge bases.
• Develop agents with deep environment understanding by integrating internal context.
• Test cloud and Kubernetes environments, modern web apps, external surface, endpoints.
• Create vulnerability-management loops including remediation tracking and fix verification.
• Design human-in-the-loop systems for dangerous actions and steer agents away from unproductive paths.
• Create feedback mechanisms for learning from decisions and domain expertise.
• Develop rigorous evaluations measuring security outcomes and agent capabilities over time.
• Build production-grade infrastructure that runs continuously with observability and safety.
• Investigate failures in agent reasoning and improve tooling, context, and guardrails.
• Partner with security and engineering teams to ensure high-signal, actionable findings.
• Help define the future of offensive security at OpenAI with scalable automation.

Skills

Offensive security

Cloud security

Kubernetes security

Web app security

Source-code review

Linux security

macOS security

External attack surface testing

Job description

OpenAI in New York seeks a Staff- Principal offensive security domain expert to architect and own production-grade security agents that continuously identify vulnerabilities, validate exploits, drive remediation, and evolve with guardrails.

You will build a portfolio of agents, integrate context, test cloud, Kubernetes, web apps, endpoints, and attack surfaces, and collaborate with security and engineering teams to deliver high-signal, actionable findings at scale.