Technical Program Manager, AI Safety and Security Launches at DeepMind leads end-to-end launch lifecycles, orchestrating global cross-functional teams to deploy Gemini models with the highest safety and security standards. This on-site role collaborates with GenAI Safety, Security, Trust and Safety, and Legal teams to manage the operational lifecycle of model deployments, streamline execution, and implement tooling upgrades across all launch programs. You will identify bottlenecks, resolve dependencies, and influence outcomes across teams without direct authority. You will author and maintain end-to-end launch playbooks, standardized templates, intake reviews, and go/no-go approval workflows. The position is based in London, United Kingdom; Mountain View, California, United States; and New York, New York, United States. US compensation is $217,000 - $236,000 (USD) plus a 15% bonus target, equity, and benefits. DeepMind emphasizes safety and ethics as part of its mission to push the boundaries of AI responsibly.

Technical Program Manager, AI Safety and Security Launches at DeepMind leads end-to-end launch lifecycles, orchestrating global cross-functional teams to deploy Gemini models with the highest safety and security standards. This on-site role collaborates with GenAI Safety, Security, Trust and Safety, and Legal teams to manage the operational lifecycle of model deployments, streamline execution, and implement tooling upgrades across all launch programs. You will identify bottlenecks, resolve dependencies, and influence outcomes across teams without direct authority. You will author and maintain end-to-end launch playbooks, standardized templates, intake reviews, and go/no-go approval workflows. The position is based in London, United Kingdom; Mountain View, California, United States; and New York, New York, United States. US compensation is $217,000 - $236,000 (USD) plus a 15% bonus target, equity, and benefits. DeepMind emphasizes safety and ethics as part of its mission to push the boundaries of AI responsibly.