I'm a KAIST undergraduate in Computer Science and Business & Technology Management, interested in how people make sense of complex information, coordinate with others, and work with AI.
I build interfaces to explore those interactions. When the thing answering you is a model, a wall of text isn't really an interface. At the KAIST AI Experience Lab I work on generative UI and human-agent interaction. Lately the same questions have pulled me toward organizations, where information, decisions, and work are spread across people, tools, and documents, and at the Collaborative Social Technologies Lab I've joined a project that puts a team of agents to work on those documents, with people in the loop, to surface the workflows already hidden inside them. Away from work, I trained cheerleaders as ELKA's head of training, and I read whenever there's a gap in the day.
Selected work
View archiveAn ongoing inspection tool for Notion, Figma, and Slack extraction. It keeps tool inputs, raw responses, failures, and omissions visible so the path from source material to structured knowledge can be checked. Built-in examples support local exploration; workflow discovery and its evaluation remain a separate research step.
Developed through KOI and continued as an open-source GenUI environment. A user's interaction trace becomes material for the next composed interface: pinning, hiding, and reordering record explicit choices that the next composition can preserve. Built on the MCP gateway packages.
Rendered LLM conversation history as a navigable constellation in WebGL, so branching threads read as spatial structure instead of scrollback. Each exchange becomes a node you can orbit, and forks stay visible as separate arms rather than collapsing into one linear log.
Skills
DESIGN
Figma MCP Agent Tooling
WEB BUILDING
TypeScript JavaScript React HTML CSS WebGL Three.js
APPS & SYSTEMS BUILDING
Flutter Swift Xcode C++ Python
SHIP & DOCUMENT
Git GitHub Typst
