Model Policy Manager, Multimodal Safety
Core
Design and maintain safety policies for multimodal AI models (audio, image, video, omni-modal) to mitigate misuse and ensure alignment with human values.
Role type
Manager, AI Safety Policy
Builds
Behavioral safety policies, evaluation criteria, grading guidance, and safeguards for frontier multimodal models.
Domain
Artificial Intelligence / Multimodal Safety
Deliverable
production ML models
Required skills
Judgment on real-world risks of advanced AI, translating threat models into measurable policies, policy iteration based on failure analysis, designing scalable policy systems, technical fluency with AI tools, hands-on data inspection and evaluation, driving consensus in ambiguous spaces.
Preferred skills
Experience with multimodal model data, pragmatic risk reduction approach, fast-paced research environment adaptability.
Technologies
AI tools for policy development and evaluation
Responsibilities
Design policies for audio, image, video, and omni-modal behavior; translate theories of harm into evaluation criteria; analyze safety regressions to inform policy updates; develop artifacts for model training and deployment; partner with researchers to operationalize policies.
Seniority
Manager, hands-on IC