Forward Deployed Engineer (FDE), Cyber - NYC

Job description

About the team OpenAI’s Forward Deployed Engineering team builds with customers to bring frontier AI into production. Our Cyber FDEs partner with the security and engineering teams that defend globally critical systems, from enterprise software to the core infrastructure it runs on. Cyber FDEs work closely with our Product, Research, and Go-to-Market teams to adapt our products and models to what customers need to defend their systems. About the role You’ll build systems that give defenders an advantage at a scale few engineering teams encounter. You might help an enterprise identify, test, and deploy patches across hundreds of thousands of code repositories, or help a security team analyze petabytes of logs for signs of malicious agents or coordinated attacks. Across these deployments, you’ll build tools that customer engineers can trust to support consequential security decisions. You’ll work directly with customer security teams, from CISOs to frontline analysts. As AI changes how attackers operate, you’ll help these teams determine where their defenses need to evolve, prioritize what to build, and validate solutions with the engineers who will operate them. Alongside OpenAI’s account teams, you’ll translate security priorities into deployed systems that help defenders cover more ground. You’ll also help shape what OpenAI builds next. When you identify a gap between what defenders need and what today’s models or products can do, you’ll partner with Research and Product to define, build, and test capabilities that address it. This role is based in New York. We use a hybrid work model of 3 days in the office per week. We offer relocation assistance. Travel up to 50% is required. In this role you will: Partner with customer security leaders and OpenAI’s Go-to-Market team to scope engagements, establish responsibilities, and define measurable success criteria. Lead technical design and implementation alongside customer engineers. Integrate agents into security workflows, establish access boundaries, and determine which actions require human review. Build evaluations grounded in the decisions defenders make. For a patching agent, test whether a fix blocks a known exploit while preserving expected application behavior. For a detection system, replay historical attacks alongside normal traffic to measure detection speed and false-positive rates. Guide productio…