Senior Offensive AI Security Engineer

This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Senior Offensive AI Security Engineer based in the United States. This role sits at the intersection of offensive security, vulnerability research, and applied artificial intelligence. You’ll investigate complex products, applications, services, and infrastructure to uncover high-impact vulnerabilities and attack paths. The position combines deep security expertise with frontier and open-weight AI models, agents, and custom research tooling. You’ll independently develop hypotheses, validate exploitability, and turn promising findings into defensible security evidence. The work is highly autonomous, research-driven, and focused on risks that conventional security testing may overlook. You’ll collaborate with engineering and security teams to communicate findings, support remediation, and verify fixes. The role also offers an opportunity to shape how AI-native security research is conducted across a broad security organization. Accountabilities: - Conduct vulnerability research across applications, products, services, and infrastructure, focusing on subtle, high-impact vulnerabilities and attack paths that cross component or trust boundaries. - Use threat analysis, architecture knowledge, source code, observed system behavior, and target familiarity to select research areas and guide investigations. - Apply frontier and open-weight AI models, agents, and other AI capabilities throughout the research lifecycle, including reconnaissance, code analysis, hypothesis generation, exploit development, and verification.

- Design and optimize research harnesses that provide AI models with appropriate context, tools, execution environments, and feedback. - Develop custom security tooling such as analysis utilities, fuzzers, agents, test harnesses, proofs of concept, and full exploits when required to answer research questions. - Validate model-generated findings by reproducing issues, eliminating false positives, establishing prerequisites, and distinguishing theoretical weaknesses from demonstrated vulnerabilities with meaningful impact. - Improve the reliability and scalability of AI-assisted security research by addressing false positives, false negatives, context limitations, nondeterministic behavior, and reproducibility challenges. - Evaluate trade-offs between constrained, orchestrated research pipelines and more autonomous tool-using agents depending on the research objective. - Partner with engineering and product security teams to communicate findings, support remediation, identify related risks, and validate fixes. - Share tools, techniques, research methods, and lessons learned to help strengthen AI-enabled security practices across the broader security organization. Requirements: - 5+ years of hands-on experience in vulnerability research, offensive security, application security, penetration testing, or a related discipline. - Demonstrated ability to select research targets, formulate and refine hypotheses, and establish exploitability and security impact in complex software or production environments. - Hands-on experience applying frontier and open-weight AI models to offensive security workflows, including selecting models appropriately when refusal behavior can interfere with legitimate security research. - Experience evaluating AI-driven security research processes, including identifying false positives, missed vulnerabilities, unstable results, and reproducibility gaps. - Strong understanding of agent architecture trade-offs and the ability to determine when controlled, reproducible workflows are appropriate versus when more open-ended agents provide value.