Hanyang Zhao

I am a fourth(final) year Ph.D. candidate at the Department of IEOR of Columbia University advised by Professor Wenpin Tang and Professor David D. Yao. Prior to my Ph.D. study, I obtained my B.S. degree in Mathematics at Fudan University and my M.S. degree in Financial Engineering also at Columbia. I also spent two wonderful summer doing research internships at Netflix (2025) and Capital One (2024).

My research has mainly focused on reinforcement learning (RL) and generative models (both LLMs and diffusion model). I try to enhance the design space of algorithms from first mathematical principles and by leveraging the structural properties of the underlying models. Some of my prior works on both theoretical and practical sides include:

(1) Efficient and effective off-policy RL for enhancing dLLMs reasoning [DiFFPO];
(2) RL in continuous time and space theory for policy optimization [CT-PPO] and applications on fine tuning diffusion models [Score as Action];
(3) Generalized preference modeling/optimization using Mallows-ranking model beyond Bradley-Terry [MallowsPO];
(4) Unified perspective on the design space of offline RLHF algorithms [RainbowPO].
(5) Noise schedule design and convergence analysis of diffusion models: [CDPM],[Tutorials].

I am now on the industrial job market! I am looking for researcher opportunities working on AI. If you think that I am a potential good fit, please email me at hz2684 AT columbia DOT edu.