Zheng Han is a graduate of the class of 2021 in Mechanical Engineering at Zhejiang University–University of Illinois Urbana-Champaign Institute (ZJUI), and currently a doctoral candidate at the Massachusetts Institute of Technology (MIT). His academic journey from ZJUI to MIT has been forged through rigorous research training, vibrant intellectual exchange, and unwavering curiosity. Today, he continues to break new ground at the intersection of large language models and intelligent systems decision-making, driven by a core mission to translate lab-based research into real-world impact.
▲
Looking back on his undergraduate years at ZJUI, Zheng Han attributes his foundational research mindset to the Institute’s interdisciplinary education model and deeply international academic environment. During his undergraduate studies, he joined the research group of ZJUI Associate Professor Cui Jiahuan, where he participated in computational fluid dynamics projects. This experience marked his first deep dive into the true nature of academic research: a rigorous process of identifying impactful questions, conducting systematic investigations, and refining ideas through iterative testing, a stark contrast to the fragmented absorption of textbook knowledge.
He also vividly recalls that many ZJUI courses centered on comprehensive, project-based learning, rather than relying solely on routine assignments. Students were challenged to conduct independent literature reviews, design methodological frameworks, execute experiments, and build fully functional systems from the ground up. This hands-on training fostered in him a habit of independent problem-solving, a skill that would prove indispensable throughout his doctoral studies.
▲
Upon enrolling at MIT in 2021, Zheng Han shifted his research focus to decision-making in intelligent systems. For him, the steepest learning curve was not language adaptation, but the transition from structured undergraduate coursework to open-ended academic inquiry. Unlike undergraduate study, where goals and milestones are clearly defined, research rarely comes with standard answers, even the most promising hypotheses can fail to bear out in experimentation. He soon found that regular, in-depth discussions with his supervisors and peers were invaluable for refining his thinking and unlocking new research directions. Over time, he came to view research not as a solitary pursuit, but as an iterative cycle of discussion, experimentation, and refinement.
▲
Today, his core work centers on the post-training of large language models and reinforcement learning, with a particular focus on their application in complex real-world scenarios involving multi-turn interaction, task planning, and tool use. His work is anchored in a core goal: to build more stable, reliable, and practically useful intelligent systems. Through collaborative industrial projects, including work on automated warehouse robotics systems, he has gained a profound understanding of the critical gap between controlled lab results and real-world field deployment. For him, truly impactful research is not only about pushing the boundaries of model performance, but about developing technologies that operate reliably in real-world environments and deliver tangible, practical value.
▲
Years after his graduation, Zheng Han has kept a close eye on the growth of ZJUI. He feels an immense sense of pride watching the Institute continue to make groundbreaking strides in research innovation and international collaboration, and seeing each new cohort of ZJUI students achieve remarkable feats across diverse industries and sectors.
Over the past decade, ZJUI has remained unwavering in its commitment to cultivating top-tier engineering talents with a global perspective, innovative capabilities, and a strong sense of social responsibility, building a bridge for countless students to the world’s leading academic arenas. And the young minds shaped here will continue to make their voices heard as proud ZJUIers, across the global frontiers of academic research and industrial development.
