Drive end-to-end program management and evolution of Google’s Frontier Safety Framework, balancing innovation with risk governance.
Scale and implement unified safety gates across Google DeepMind, Research, and product pipelines for frontier model training and deployment while acting as the central interface for leadership updates on model risk posture, and collaborate with external policy, academic, and industry partners to align on global safety standards.
Coordinate "early warning evaluations" to monitor Critical Capability Levels across core domains (autonomy, cyber, biosecurity, etc).
Manage rapid mitigation plans for model CCL breaches, coordinating weight exfiltration security and misuse-limiting deployment protocols.
Serve as the central liaison for leadership updates on model risks, collaborating with external policy, academic, and industry partners on global safety standards.

Technical Program Manager, Frontier Safety, Alignment and Collaboration, DeepMind

Job Description