Tianyu Wang
 |
Postdoctoral Fellow
Email: wangtianyu6162 [AT] gmail [DOT] com To be uncertain is uncomfortable, but to be certain is ridiculous. -- Goethe |
Research Overview
Trustworthy decisions: I develop diagnostic tools and trustworthy methods across the data-to-decision pipeline, focusing on often-overlooked yet practically important components, including evaluation criteria [1] and data acquisition.
AI evaluation: I study systematic approaches to improving modern AI evaluation from an operations research perspective, including benchmark design [2] and calibration.
Healthcare operations: I partner with companies and government agencies to deploy AI-driven algorithms in healthcare settings [3].
Bio
I am a postdoctoral fellow at Harvard Business School AI Institute and will join The Chinese University of Hong Kong (CUHK) as an Assistant Professor in the Department of Systems Engineering and Engineering Management (SEEM) in July 2027.
I study how to design, evaluate and deploy AI-driven decision making, with a recent focus on AI evaluation and healthcare optimization. My research lies at the interface of operations research, statistics, and machine learning. I have been collaborating with industry partners to implement data-driven optimization models in real-world operations, including the Mount Sinai Hospital and the Fire Department of the City of New York.
My work has been published in leading business journals, including Management Science and M&SOM, as well as major AI conferences, including NeurIPS and ICML. My research has received recognitions from INFORMS George B. Dantzig Dissertation Award [25-page thesis summary], George Nicholson Student Paper Competition, INFORMS Applied Probability Society, Stochastic Programming Society, among many others. Previously, I completed my PhD in Operations Research from Columbia University, where I was fortunate to be co-advised by Professors Garud Iyengar and Henry Lam. Before that, I obtained my BE in Information Systems and BS in Mathematics from Tsinghua University in 2021.Featured Research
[1] Optimizer's Information Criterion: Dissecting and Correcting Bias in Data-Driven Optimization. Management Science (minor revision).
[2] Rethinking Distribution Shifts: Empirical Analysis and Modeling for Tabular Data. Management Science. Forthcoming.
[3] Translating Time-Series Foundation Models into AI-Enhanced Medication Inventory Control in a Large Urban Health System. Working Paper.
Check out the By Topic page for an overview of current research directions, and the Research page for a full list of working and published papers.