I am an Advanced AI Research Scientist Manager at the Accenture Center for Advanced AI, where I work on self-evolving LLM agent training and optimization methods for improving reasoning capabilities.

Previously, I was a Postdoctoral Research Scientist at Columbia University, hosted by Prof. Samory Kpotufe, working on the theory of transfer learning and multi-task learning.

I received my Ph.D. in Computer Science and Engineering from The Pennsylvania State University in 2024, advised by Prof. Mehrdad Mahdavi. Before that, I earned my B.E. in Communication Engineering from Soochow University in 2018.

My research focuses on machine learning from heterogeneous data sources, with interests spanning:

  • LLM post-training and self-improving learning
  • Transfer learning, multi-task learning and federated learning
  • Transferrable Agent Evaluation Method
  • Large-scale, minimax and non-convex optimization

News

  • 2026: Paper on distributed sub-model training accepted to AISTATS 2026.
  • 2026: Paper on Neyman-Pearson classification under distribution shift accepted to ICLR 2026.
  • Jan. 2026: Joined Accenture Center for Advanced AI as an Advanced AI Research Scientist Manager.
  • 2025: Paper on Mixed-Sample SGD accepted to NeurIPS 2025.
  • 2025: Paper on stochastic compositional minimax optimization accepted to AISTATS 2025.
  • Sep. 2024: Joined Columbia University as a Postdoctoral Research Scientist.
  • Aug. 2024: Defended my Ph.D. dissertation at Penn State.

I maintain a publication list on Google Scholar.