UIST 2026
CaseSense: Mapping the Legal Case Landscape with Multi-Dimensional Visual Exploration
Multi-dimensional visual exploration of legal case landscapes
Junghwan Kim김정환
👋 HAI & LLM Researcher · Ph.D. Candidate at Seoul National University
I am a Human-AI Interaction (HAI) and LLM researcher, and a Ph.D. candidate in the
Department of Intelligence and Information at
Seoul National University,
advised by Bongwon Suh in the
Human-Centered Computing Lab.
I study human-centered interactive systems.
Models have improved fast, but the benefit people feel has not. The tacit knowledge experts build over years, what they judge on and how, still has to pass through a single chat box. My research externalizes that knowledge into the structure of interactive systems: I find problems in fields such as law and medicine, build the full stack from data pipelines and agent orchestration to the interactive frontend, and derive design insights from evaluations with real users.
As generation gets easier and human judgment matters more, my goal is a generalized set of guidelines for human-centered AI system design.
Research Highlights ✨view all 17 →
Interactive systems that externalize expert tacit knowledge, at the intersection of human-AI interaction and LLM systems.
ITS 2026
🏆DEBSIM: Productive Friction for Learning Transfer in LLM-Based Argumentation Tutoring
Productive friction for learning transfer in LLM argumentation tutoring
Best Full Paper Award
CogSci 2026
🏆Co-Overcooked: Cognitive Constraints on Partner Modeling in Human-AI Team Composition
Cognitive constraints on partner modeling in human-AI teams
Student Travel Award
UIST 2026 Adjunct
ToneCanvas: Visually Painting Character Tone for Narrative Consistency
Persistent watercolor overlays for comparing and revising character tone across a manuscript
CSCW 2026 Companion
Between the Clinic and the Kitchen: A Constraint RAG Chatbot as Boundary Infrastructure for Dialysis Care
A constraint RAG chatbot as boundary infrastructure for dialysis care
CHI 2026 EA
When Nurses Meet Vibe-Coding: Shifting from Problem Talk to Situated Specification in Participatory Design
Real-time LLM prototyping shifts participatory design toward situated specification
UIST 2025
LegisFlow: Enhancing Korean Legal Research with Temporal-Aware LLM Interfaces
Temporal-aware LLM interfaces for Korean statutory research
📰News
Sep 2026
Yulchon AI Young Researcher Scholarship
Selected for the Yulchon AI Young Researcher Scholarship (AI+X) by the Nongshim Yulchon Foundation and the SNU AI Institute.
Aug 2026
Journal article accepted at the Interactive Journal of Medical Research
Our mixed-methods RCT of a personalized RAG chatbot for chronic kidney disease patients was accepted at i-JMR.
Jul 2026
Co-Overcooked presented at CogSci 2026 in Rio de Janeiro
Oral presentation at CogSci 2026, supported by a Student Travel Award.
Jun 2026
Best Full Paper Award at ITS 2026
DEBSIM received the Best Full Paper Award at the International Conference on Intelligent Tutoring Systems in Pafos, Cyprus.
Apr 2026
Two extended abstracts presented at CHI 2026 in Barcelona
When Nurses Meet Vibe-Coding, and From Objects to Influences.
Feb 2026
Two extended abstracts accepted to CHI 2026
Papers on vibe-coding in participatory design and on deletion in learning-based systems were accepted to CHI EA 2026.
Oct 2025
LegisFlow presented at UIST 2025 in Busan
Presented our temporal-aware LLM interface for Korean legal research.
Aug 2025
LegisFlow accepted to UIST 2025
Accepted as a full paper at the 38th ACM Symposium on User Interface Software and Technology.
🗓️Upcoming
🏅Honors
- Sep 2026Yulchon AI Young Researcher Scholarship (AI+X)
- Jun 2026Best Full Paper Award
- Jul 2026Student Travel Award
🔍Interests
🤝 Human-AI Interaction🧩 LLM-Augmented Interfaces⚖️ Legal Tech🩺 Health Informatics📚 Education✍️ Writing & Translation