Mohsen Bayati

Timeline

1. 1997–2007

   Mathematics and electrical engineering

   He wins a gold medal at the International Mathematics Olympiad in 1997. After a BS in mathematics at Sharif University of Technology, he enters Stanford as a mathematics PhD student in 2000 and moves to its electrical engineering department. He completes an electrical engineering doctorate in 2007, with a dissertation on algorithms and models for large-scale networks. He also earns an MS and completes a PhD minor in mathematics at Stanford University; the dates are unknown.

2. 2007–2011

   Postdoctoral research on machine learning

   From 2007 to 2009, he works as a postdoctoral researcher at Microsoft Research on machine learning and optimization for healthcare and online advertising. He holds a postdoctoral position in Stanford's Electrical Engineering department from 2009 to 2011, studying high-dimensional statistical learning.

3. 2011–2015

   Joining Stanford business faculty

   In 2011, he joins Stanford's faculty. He becomes an associate professor of Operations, Information, and Technology at Stanford Graduate School of Business in 2015.

4. 2014

   Predicting heart failure readmissions

   In 2014, he coauthors Data-Driven Decisions for Reducing Readmissions for Heart Failure: General Methodology and Case Study. The team builds a statistical classifier from 793 hospital visits to predict readmission within 30 days of discharge, then combines predictions with decision analysis to guide care after discharge.

5. 2014–2016

   INFORMS and NSF awards

   His work receives the INFORMS Health Applications Society's William Pierskalla best paper award in 2014 and 2016, and the INFORMS Applied Probability Society best paper award in 2015. He receives a National Science Foundation CAREER Award in 2016.

6. 2019–2026

   Roles at Amazon and Turing

   From 2019 to 2025, he serves as an Amazon Scholar, working on pharmacy safety, language models and experimentation. He serves as a Distinguished Professor at Turing from 2025 to 2026, working on the evaluation of AI and agentic systems.

7. April 2024

   MEDIC for medication directions

   In April 2024, he coauthors Large language models for preventing medication direction errors in online pharmacies, introducing MEDIC. The system combines a language model, pharmacist knowledge and safety rules to generate medication directions; during a deployment experiment in an online pharmacy, it reduces errors caught before reaching patients by 33%.

8. October 2025

   The Oversight Game

   In October 2025, he and William Overman submit The Oversight Game: Learning to Cooperatively Balance an AI Agent's Safety and Autonomy. They model control as a two-player Markov game in which an AI agent chooses whether to act or ask, while a human chooses whether to trust or oversee it.

9. May 2026

   Calibrated Collective Oversight

   In May 2026, he and William Overman introduce Calibrated Collective Oversight in Calibrating Conservatism for Scalable Oversight. The method combines several scoring functions into a penalty for departures from a conservative baseline and adjusts the penalty using Conformal Decision Theory.

10. As of October 2026

    Stanford professor and Google researcher

    As of October 2026, he is the Carl and Marilynn Thoma Professor of Operations, Information and Technology at Stanford Graduate School of Business, with courtesy appointments in Electrical Engineering and Radiation Oncology. He works on AI alignment, human oversight and sequential decision-making. In 2026, he becomes a visiting faculty researcher at Google.