Sadegh Shirani
Embedded Files
I am an Assistant Professor of Operations Management at the MIT Sloan School of Management.
I received my PhD in Operations, Information & Technology from the Stanford Graduate School of Business, where I was fortunate to be advised by Mohsen Bayati.
My research develops principled methods for learning, experimentation, and decision-making in complex systems. A central theme of my work is the two-way relationship between AI and experimentation: using AI to improve experimentation and using experimentation to design and evaluate AI systems. In one direction, I study how AI models and methods can improve the design, analysis, and interpretation of experiments, particularly when the available data are complex or unstructured. In the other, I use causal inference and experimental design to study the behavior and safety of AI systems at scale, especially in interactive and multi-agent settings. I also develop LLM-based simulations of human and human–AI interactions as controlled environments for experimentation. The common challenge across these problems is to draw reliable causal conclusions and make sound decisions when interactions are complex, important features are only partially observed, and data are limited. I study the theoretical foundations of these problems and develop methods that are both theoretically grounded and practically reliable.
My work combines ideas from probabilistic modeling, statistical physics, and optimization. My research is mainly motivated by systems where uncertainty and limited data make reliable decision-making challenging. Examples include public health interventions, experimentation on online platforms, and interactive AI systems.
I am always happy to discuss research ideas and potential collaborations with students and colleagues, and I am actively looking for students interested in these research directions.
Email: sshirani 'at' mit 'dot' edu
News/Highlights
News/Highlights
January 2026: Stanford GSB “Insights” covered our work LLM-SocioPol, as an AI-based virtual world designed to run controlled experiments when real-world experimentation is limited.
October 2025: Launched LLM-SocioPol: an open-source Python code for simulating LLM-based social networks to study causal effects of social influence on voting behavior, with data from five independent runs included.
August 2025: Featured in Stanford GSB “Voices” — sharing my journey as a PhD student in Operations, Information, and Technology.
June 2025: Awarded the Gerald J. Lieberman Award — a distinguished Stanford doctoral fellowship, recognizing outstanding achievements and strong potential for leadership in academia.
June 2025: Launched CausalMP: an open-source Python package for estimating causal effects under network interference, with six semi-synthetic experimental environments; data from ten independent runs included.
Working Papers
Working Papers
(1) Causal Effects with Unobserved Unit Types in Interacting Human–AI Systems
with W. Overman, M. Bayati, 2026.
ICML workshop on Technical AI Governance Research, 2026
(2) Can We Validate Counterfactual Estimations in the Presence of General Network Interference? [Codes]
with Y. Luo, W. Overman, R. Xiong, M. Bayati, Major Revision at Management Science, 2026.
Finalist, INFORMS Pierskalla Best Paper Competition, 2026
Second place, INFORMS Revenue Management and Pricing (RMP) Jeff McGill Student Paper Award, 2025
Honorable mention, INFORMS Health Applications Society (HAS) Best Student Paper Competition, 2025
Accepted for oral presentation at the Conference on Digital Experimentation @ MIT, 2025
Accepted for presentation at the MSOM Technology, Innovation, and Entrepreneurship SIG, 2025
(3) Validating Causal Message Passing Against Network-Aware Methods on Real Experiments
with A. Tan, J. Nordlund, M. Bayati, 2026.
(4) On Evolution-Based Models for Experimentation Under Interference
with M. Bayati, 2025.
(5) Simulating and Experimenting with Social Media Mobilization Using LLM Agents [Codes]
with M. Bayati, 2025.
(6) Asymptotic Analysis of Multi-Class Advance Patient Scheduling
with H. Abouee-Mehrizi and M. K. S. Faradonbeh, Major Revision at Management Science, 2025.
Second place, INFORMS Health Applications Society (HAS) Best Student Paper Competition, 2023
Finalist, Canadian Operations Research Society Student Paper Award, 2022
Accepted for presentation at the MSOM Healthcare SIG, 2023
Journal Papers
Journal Papers
(1) Causal Message Passing for Experiments with Unknown and General Network Interference [Codes]
with M. Bayati, Proceedings of the National Academy of Sciences (PNAS) 121(40), 2024.
Honorable mention, George Nicholson Student Paper Competition, 2024
Finalist, MSOM Student Paper Competition, 2024
Oral presentation at the Conference on Digital Experimentation @ MIT, 2024
(2) Departure Time Choice Models in Urban Transportation Systems Based on Mean Field Games
with M. Ameli, JP Lebacque, H. Abouee-Mehrizi, L. Leclercq, Transportation Science 56(6):1483-1504, 2022.
Best Paper Award, INFORMS TSL Urban Transportation SIG, 2022
Refereed Conference Papers
Refereed Conference Papers
(1) Higher-Order Causal Message Passing for Experimentation Under Unknown Interference
with M. Bayati, Yuwei Luo, William Overman, Ruoxuan Xiong, Neural Information Processing Systems (NeurIPS), 2024.
(2) Online Reinforcement Learning in Stochastic Continuous-Time Systems
with M. K. S. Faradonbeh, Proceedings of Thirty Sixth Conference on Learning Theory (COLT), PMLR 195:612-656, 2023.
(3) Thompson Sampling Efficiently Learns to Control Diffusion Processes
with M. K. S. Faradonbeh and M. Bayati, Neural Information Processing Systems (NeurIPS), 2022.
(4) Bayesian Algorithms Learn to Stabilize Unknown Continuous-Time Systems
with M. K. S. Faradonbeh, IFAC International Workshop on Adaptive and Learning Control Systems (ALCOS), 2022
(5) Mean Field Games Framework to Departure Time Choice Equilibrium in Urban Traffic Networks
with M. Ameli, JP. Lebacque, H. Abouee-Mehrizi, L. Leclercq, Transportation Research Board 100th Annual Meeting (TRB), 2021
Page updated
Google Sites
Report abuse