Fig. 00 — Introduction

Yunha
Lee

I build AI agents that operate expert software — for mobility, emergency response, and the cities in between.

Stanford ’30TUPA Lab, KAISTDochakyunhalee@stanford.edu

Scroll

Selected work

01 / 04 · 2024 — 2026

P-01MOLIT Grand Prize, 2025

Ambulance–ER Agent Network

Ambulance, traffic, hospital and insurance agents negotiate an admission in about 50 seconds — not the nearest hospital, but the nearest one that can take the patient.

P-02ISEF 2025 · HP Innovator Award

Smart crosswalk signal

YOLOv11 on crosswalk CCTV estimates each pedestrian’s speed and adapts the signal: longer for slow walkers, shorter when the crossing is empty, a warning when someone steps out on red.

P-03Dochak, 2026

Dochak Prism

A web dashboard for PTV VISSIM where an agent edits and runs simulations in plain language — and every edit passes a validation layer and one human approval before it touches the model.

P-04KAIST, 2026

TUPA Lab website

The site for KAIST’s Transport and Urban Planning with AI lab: Next.js on Vercel, content pulled hourly from Notion tables so a non-developer runs it, with a built-in fallback and full Korean–English parity.

~50s

From intake to a confirmed emergency room to insurance, in one run of the ambulance agent network.

agent types across institutions

live APIs in the demo run

94%

of paramedics say ER dashboards need replacing

What I’m working on

Now

Right now I’m working on how AI agents in the physical world find and trust agents they were never programmed to know. Robots, cars, and city systems built by different companies will need to work together in seconds, and today they mostly can’t. Longer term, I want to take this beyond Earth. A message from Mars takes minutes to reach us, so robots there will have to find each other and decide together on their own.

About

이윤하

I’m a freshman at Stanford. Before that I spent a summer at KAIST’s TUPA lab and at Dochak, a KAIST-spinoff mobility startup, building the same kind of thing in two settings: AI agents that drive software experts already use, with the guardrails that make an expert trust them.

Experience

Stanford UniversityClass of 2030 · from Sep 2026
TUPA Lab, KAISTResearch Intern, Transport and Urban Planning with AI · Prof. Inhi Kim · Summer 2025

Awards

01Minister’s Award (1st place)

02Regeneron ISEF 2025 Special Award: HP Innovator Award

Beyond work