Ziyu He (何子彧)
Postdoctoral Research Associate
Department of Data Science and Operations
Randall R. Kendrick Global Supply Chain Institute
Marshall School of Business, University of Southern California
Postdoctoral Research Associate
Department of Data Science and Operations
Randall R. Kendrick Global Supply Chain Institute
Marshall School of Business, University of Southern California
I am currently a post-doctoral research associate in the department of Data Sciences and Operations at the Marshall School of Business, University of Southern California (USC), where I am also affiliated with the Randall R. Kendrick Global Supply Chain Institute, advised by Professor Vishal Gupta and Professor Nick Vyas. I obtained my Ph.D. degree from the Daniel J. Epstein Department of Industrial and Systems Engineering at USC, where I was advised by Professor Jong-Shi Pang.
My research focuses on data-driven optimization for emerging decision-making environments, particularly those arising from supply chain digitization, where two major trends stand out: (1) increasing operational complexity from technologies like AI and 3D printing, demanding more advanced decision models; and (2) the explosion of data, requiring scalable machine learning algorithms. At their intersection lies optimization, and my overarching goal is to develop optimization theories and algorithms tailored to special structures embedded in these environments to enable efficient, provably effective decision support. Specifically, my research centers on two complementary themes:
Efficient large-scale optimization methods, especially for models that are nonconvex and/or under uncertainty
Modern machine learning and operations increasingly involve nonconvex optimization under uncertainty. Going beyond classical toolkits, my research designs algorithms that better balance solution quality and computational efficiency through structures like difference-of-convexity, piecewise smoothness and submodularity—common in applications such as deep learning, sparse learning, and Bayesian inference.
Facilitating adoption of new technologies, especially those that require major restructuring in operations
Technological innovations like 3D printing offer strategic potential for efficiency, flexibility and resilience, but require substantial operational restructuring—making adoption decisions complex and uncertain. My research builds quantitative models that clarify when, why, and how such technologies generate value, and helps organizations assess potential gains using their own data.
To advance these research themes, I also work closely with industry collaborators to ground my research in real-world challenges and data, where the analytical insights and data-driven tools drawn from my research directly support their decision-making. Currently, I am collaborating with Kwai, a leading digital content platform expanding into e-commerce, on operational policies that balance user retention with revenue and manage AI-generated content.
If you're interested in collaborating on similar research or simply want to grab a coffee, please feel free to reach out at ziyuhe@usc.edu.