Hi, I’m Madhu. I head the AI Safety Program at the Partnership on AI and serve on the Advisory Board of the Centre for the Study of Existential Risk at the University of Cambridge, where I’m also a Research Affiliate. I focus on emerging technologies and their policy implications to drive pragmatic, collective oversight.

My work spans policy research, strategic foresight, and research program leadership— aimed at shaping industry practices and informing regulation. I’m passionate about building collaborations across industry, government, and academia to drive consensus on urgent policy questions and turn ethical principles into concrete action.

Previously, I was a Public Interest Technology Fellow at New America and an Associate Fellow at the Observer Research Foundation in New Delhi. My research has been featured in Nature Machine Intelligence and supported by the Ford and Hewlett Foundations, and cited by the U.S. AI Safety Institute. I hold an LL.M. from Harvard Law School, where I was an Inlaks Scholar. I live in San Francisco.

Recent Work:

– Led development of PAI’s Guidance for Safe Foundation Model Deployment, a framework co-created with 40+ organizations and endorsed by Meta, Microsoft, and Google to support scalable safety practices based on model capabilities and release types.

– Currently leading a $700,000+ research grant from Georgetown’s CSET to create voluntary industry guidelines for monitoring AI agents.

– Led primary research on foreign law enforcement access to digital evidence, providing policy recommendations to U.S. and Indian governments.