AI Agent Abuse Prevention Engineer
Zendesk
Texas, Washington, California, Arizona, Massachusetts, South Carolina, North Carolina, Illinois, Georgia, Florida, Wisconsin
Workplace: RemoteFull timeUSD 200,000 - 300,000 annuallyFunction: Data Science & Machine LearningExperience: 10+ yearsSkills: ["Threat modeling","Technical leadership","Cross-functional collaboration","Communication","Incident investigation"]Own detection, mitigation, and governance of abuse from AI agents across Zendesk products and customer integrations. Lead threat modeling for agent features, integrations, and APIs, and design secure agent architectures with mitigation controls such as sanitization, provenance tracing, policy gates, and sandboxing. Build anomaly detection using telemetry and behavioral analytics, partner with security on containment, and serve as an SME during investigations while collaborating across Platform, Product, SDK, Legal/Privacy, and Customer Success.
