Dongyeop Kang Earns McKnight Land-Grant Professorship

Department of Computer Science & Engineering Assistant Professor Dongyeop Kang has earned the McKnight Land-Grant Professorship for his work developing human-centered AI systems that support expert cognitive workflows. The two-year grant, $50K program aims to advance the careers of new assistant professors at a crucial point in their professional lives.

"When I first joined the University of Minnesota in 2021, the vision I had for this field was nowhere near the mainstream of Natural Language Processing (NLP) research," Kang said. "Convincing other AI researchers why human-centered AI matters — why it goes beyond simple scaling — was never easy. And because these are problems that cannot be solved without interdisciplinary collaboration, building the lab and training students in those early years was genuinely hard. Now, five years in, preliminary results are starting to emerge, and I can feel the field beginning to recognize the importance of this direction. To have University also recognize the value of my research vision means a great deal to me. My original plan was a 10-year journey, and this award gives me real confidence that the foundation we've built is worth expanding."

Kang's research group, Minnesota NLP, is focused on building human-centric language technologies, cognitively aligning human and machine thinking, and advancing AI as “thinking partners”. His research begins where today's AI falls short. Traditional NLP has focused on building models that understand and generate text, but these systems still fail to engage with the processes of human thought. His work bridges language and cognition to develop novel algorithms, cognitive metrics, and interaction frameworks that assist experts at real-world workplaces in their complex cognitive workflows.

At the heart of Kang's vision is what he calls “human-centered amplification” — designing AI that does not replace human reasoning, but augments it. Rather than automating expert tasks, his systems learn from how people plan, reason, and create; anticipate cognitive bottlenecks; scaffold difficult tasks; and adapt dynamically to expert strategies. The goal is building AI that understands cognitive intent, supports deep analytical thinking, and collaborates as a true partner in expert judgment.

"Among all our projects, the ones I find most ‘challenging’ and most rewarding are those where we collect months of real workflow data from knowledge workers in specific domains, like lawyers and scientists," Kang said. "Gathering two weeks to six months of their thinking processes and final outputs, comparing their thinking to AI's, and understanding how AI can truly support human cognitive states — that is the foundational work everything else is built on."

Among the lab's active projects is LawFlow (COLM 2025), developed in collaboration with the UMN Law School and supported by Open Philanthropy. The project maps the differences between how human and AI lawyers reason through complex cases, laying the groundwork for next-generation AI that professionals can communicate and genuinely collaborate with. 

In parallel, ScholaWrite (under review, ACL 2026), supported by Grammarly and conducted in collaboration with UMN Educational Psychology, traces the end-to-end research and writing process of scientists across months-long cycles, capturing fine-grained signals such as conversation logs and writing keystroke data. Scholarly writing is a cognitively complex demanding task that takes three to six months and requires multiple scientists to weave fragmented knowledge into a coherent story that persuades fellow researchers. Kang's vision is to extend this work from writing into the full research workflow including coding, communication, and planning, and eventually into physical sciences, such as chemistry and biology toward self-driving labs, building longitudinal, personalized, self-evolving research partners that can accelerate scientific discovery.

Kang joined the Department of Computer Science & Engineering in 2021 as an assistant professor, where he founded the Minnesota NLP group — the first of its kind in the department. He received his BS/MS (2010) in computer science engineering from the Korea Advanced Institute of Science and Technology (KAIST) and his PhD (2020) in computer science from Carnegie Mellon University. Prior to joining the University, he served as a research intern at Microsoft Research, the Allen Institute for Artificial Intelligence (AI2), and Facebook AI Research in 2016, 2017, and 2018, respectively, and as a postdoctoral scholar at the University of California, Berkeley, from 2020–21. 

Together with incoming NLP faculty member Alexander Spangher, Kang continues to lead the Minnesota NLP group as an interdisciplinary hub for human-centered AI research, creating systems that reflect human values and support cognitive growth.

Learn more about Kang’s work at his personal website and the Minnesota NLP website.