My name is Haoyang (pronounced as “Hau Yeung”); I am a PhD candidate at the School of Mechanical Engineering, Purdue University, advised by Prof. Guang Lin.

My current research focuses on LLM post-training and efficient generation, with methodological experience spanning on-policy distillation, supervised fine-tuning, reinforcement learning, and sampling/optimization.

I was a Research Intern at Google, focusing on LLM post-training. Previously, I was RedStar Intern (ACE Intern Program) at Xiaohongshu Inc.’s Humane Intelligence Lab (hi lab) (Beijing, China), where I worked on Multimodal LLMs.

Research Interests

• Post-training
• LLMs
• Agents

News

• [Sep/2026] 2 papers were accepted by NeurIPS 2026
• [May/2026] Research Intern at Google, working on LLM Post-Training
• [Jan/2026] 2 papers were accepted by ICLR 2026
• [Jan/2026] Invited talk at the Crunch Seminar, Brown University
• [Jan/2026] Invited talk at the Computational Scientific Imaging Lab, Peking University
• [Dec/2025] ACE Intern Program at Xiaohongshu Inc, Humane Intelligence Lab (hi lab), working on multimodal LLMs
• [Nov/2025] 1 paper was selected for oral presentation at AAAI 2026
• [Oct/2025] 1 paper was accepted by Journal of Computational Physics
• [Sep/2025] 🔥 Feel the Ultra-Fast Language Generation in our latest work (60× faster than ARMs): Discrete Diffusion Divergence Instruct
• [Oct/2024] Student Travel Award at 2024 Mathematical and Scientific Foundations of Deep Learning Annual Meeting
• [Aug/2024] Student Travel Award at 2024 SIAM Conference on Mathematics of Data Science (MDS24)
• [Jul/2024] Invited talk at the Crunch Seminar, Brown University
• [May/2024] 1 paper was accepted by ICML 2024
• [Apr/2024] Passed the preliminary exam and became a Ph.D. candidate
• [Apr/2024] Student Travel Award at the ICERM workshop about Nonlocality and Interacting Particle Systems
• [Jan/2024] 1 paper was accepted by AISTATS 2024
• [Dec/2023] 1 paper was accepted by Journal of Computational Physics
• [Sep/2023] Student Travel Award at 2023 Mathematical and Scientific Foundations of Deep Learning Annual Meeting
• [May-Jul/2023] Givens Associates at Argonne National Laboratory
• [Sep/2022] 1 paper was accepted by NPJ Digital Medicine

Services

• Journal Reviewer: TPAMI, TMM, TNNLS, TCYB, TMLR
• Conference Reviewer: ICLR, ICML, NeurIPS, AISTATS, AAAI, KDD

Contact

• Email: lastname+528 at purdue dot edu; firstname+lastname+ai at google dot com