Senior AI Engineer, Security Infrastructure

A senior engineer on Air's Agentic AI team focused on the security of Ace, the company's AI agent used across US government, working where applied AI research, offensive security and production engineering meet. The engineer threat models agent architectures, researches and runs adversarial evaluations for prompt injection, tool abuse, privilege escalation, exfiltration and memory poisoning, builds automated security regression testing, and turns findings into production mitigations. The role also designs sandboxes for untrusted agent workloads, permission models for agent tools, and the Kubernetes and AWS infrastructure behind secure inference.

Air wants US citizenship, five or more years building production or security systems, demonstrated AI red teaming or offensive security work, and strong Python and Kubernetes experience. Offices in Pittsburgh and Arlington or remote, with up to 25 percent travel.