Sandhini Agarwal

Sandhini Agarwal is an AI policy researcher and Member of Technical Staff at OpenAI, where she has worked since January 2020 on the Deployment Planning team focused on safe and responsible model deployment. She is best known for leading the bias analysis for GPT-3, building OpenAI's first content filter, contributing to CLIP and InstructGPT, and authoring system cards for GPT-4, GPT-4o, and OpenAI o1. She holds a B.A.S. in Symbolic Systems with a concentration in AI and Philosophy from Stanford University.