Senior Security Engineer, Agentic Red Team, DeepMind

Google•Zurich, CA

20h•$174,000 - $252,000

About The Position

As a Senior Security Engineer on Google DeepMind’s Agentic Red Team, you will be a specialized, high-velocity member of DeepMind Security focused on closing the gap between emerging AI capabilities and traditional security reviews. You will serve as the primary technical executor of adversarial engagements, working directly with product builders to identify architectural flaws during design and development, before formal reviews begin. The role focuses on multi-turn attacks against production-level AI models, particularly agentic behaviors such as tool use and reasoning chains. You will help turn vulnerability findings into reusable protections by contributing to automated red-teaming frameworks and defensive strategies for Google agent developers. The posting also includes general background describing Google DeepMind’s mission and work on AI agents, including agent testing, simulations, visualizations, leaderboards, and robot algorithms. Compensation in the US is listed as $174,000–$252,000 USD, plus a 15% bonus target, equity, and benefits. Individual pay depends on job-related skills, experience, and relevant education or training.

Requirements

Bachelor’s degree in Computer Science, Information Security, or equivalent practical experience.
At least 5 years of experience in red teaming, offensive security, or adversarial machine learning.
Experience with large language model architectures, agentic workflows (e.g., chain-of-thought reasoning), and AI vulnerability classes.
Experience developing exploits for generative AI models, such as prompt injection, adversarial examples, and training-data extraction.

Nice To Haves

Experience working in a consulting capacity with product teams or in a startup-like environment.
Familiarity with AI safety benchmarks, evaluation frameworks, and fuzzing techniques.
Ability to translate complex probabilistic risks into actionable engineering fixes for developers.
Excellent coding skills in Python, Go, or C++, with experience building security tools or automation.

Responsibilities

Conduct rapid security assessments of agentic services, focusing on generative AI vulnerabilities such as prompt injection, tool-use escalation, and autonomous lateral movement.
Engineer and execute attack sequences that exploit non-deterministic model behavior, agentic logic errors, and data-poisoning vectors.
Convert manual vulnerability discoveries into automated regression-testing frameworks (“Auto Red Teaming”) to help prevent regressions in future model versions.
Partner directly with developers during design and build phases, providing immediate feedback to shorten the loop between offensive findings and defensive engineering.
Maintain and expand a library of agent-specific attack patterns and exploit primitives to establish release criteria for new models.