Mayank Agrawal

mayank at roundtable dot ai

I'm currently building the proof-of-human API, an invisible Turing Test for the Internet.

The AI revolution is ushering in a new age of authenticity. At Roundtable, we're building a digital trust layer rooted in behavior.

Fraudsters can fake credentials, spoof devices, and steal passwords. But they can’t fake behavior at scale. We build models that separate real from fake by analyzing session dynamics, cognitive patterns, and interaction flows -- verifying identity in real-time, without friction.

Previously

• studied cognitive science, decision theory, and machine learning at Princeton with Jon Cohen, Tom Griffiths, and Nathaniel Daw.
• ran cross country and track & field at Swarthmore while mapping moral epistemology onto computational learning theory
• went viral on Hacker News during YC S23
• conducted biotech ML research at Kallyope and data science ops at Mercury Fund
• taught amazing students at Breakthrough Houston
• characterized NBA team chemistry dynamics via multidimensional vector spaces while also making viral YouTube player mixtapes

Selected Publications:

• "Scaling up Psychology via Scientific Regret Minimization", PNAS, 2020
• "The Temporal Dynamics of Opportunity Costs: A Normative Account of Cognitive Fatigue and Boredom", Psychological Review, 2021
• "Using Large-Scale Experiments and Machine Learning to Discover Theories of Human Decision-Making", Science, 2021