AI Agent Abuse Prevention Engineer

Job Description

Zendesk is hiring a Senior Staff–level technical leader to own detection, mitigation, and governance of abuse stemming from AI agents across Zendesk’s products and customer integrations. You will combine hands-on offensive and defensive skills with product sensibility to design secure agent architectures, and champion company-wide and engineering changes to prevent AI agent-enabled abuse.

Key responsibilities

- Threat strategy & modeling

- Lead threat modeling for AI agent features, integrations, and APIs (prompt injection, jailbreaks, data leakage, automated workflow abuse).

- Maintain an evolving attacker playbook specific to AI agents.

- Technical leadership & engineering collaboration

- Design and prototype mitigation controls (e.g., input/output sanitization, provenance tracing, policy gates, token/session lifetimes, capability scoping, sandboxing).

- Work with Platform, Product, and SDK teams to bake secure defaults into agent SDKs, apps, and API flows.

- Detection & prevention

- Build anomaly detection rules, telemetry, and behavioral analytics to surface anomalous agent activity and abuse patterns.

- Partner with security teams to instrument key signals and automate high-confidence containment actions.

- Incident response & forensics

- Act as a subject-matter expert for investigations involving AI agents; define containment, eradication, and customer communications playbooks.

- Stakeholder engagement & external representation

- Collaborate with Legal/Privacy for compliance, Product for roadmap trade-offs, and Customer Success for mitigation support.

- Represent Zendesk in industry forums on agent safety/security and contribute to standards/best practices where appropriate.

Required qualifications

- 10+ years of professional experience in cybersecurity, software engineering, or ML security with demonstrable hands-on experience.