Senior Security Engineer, Agentic Red Team
Google DeepMindMountain View, CA / New York, NY / Zürich, Switzerland
AI safety & policy
Information security
Full-time
AI safety & policy
Information security
Full-time
Routes to impact
Direct high impact on an important cause
Skill-building & building career capital
Description
Lead adversarial engagements against production AI agents, finding security flaws before new agentic capabilities launch and turning findings into reusable guardrails.
- Run rapid security assessments targeting prompt injection, tool-use escalation and autonomous lateral movement
- Engineer multi-turn attacks exploiting agentic logic errors, reasoning chains and data poisoning vectors
- Turn manual findings into automated “Auto Red Teaming” regression testing frameworks
- Partner with developers during design and build phases to give immediate security feedback
- Maintain a library of agent-specific attack patterns used to set release criteria for new models
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.



