AI engineer, PhD · St. Gallen, Switzerland
Hi, this is Hangyeol!
I build ML systems that end up in production: on a factory floor, in a trading workflow, inside a robot.
An inspection system I built from scratch runs every day at a semiconductor-parts maker, cutting inspection time by over 90%. My LLM pipelines valued the options hidden in trading contracts at a commodity trading company. The agentic LLM platform I designed for the humanoid robot Nadine ran live at 12 international events. I am finishing a PhD in computer science at the University of Geneva.
now: finishing my PhD in November 2026, moving to St. Gallen and available from January 2027.
The LLM brain of the humanoid robot Nadine: a ten-agent LangGraph system and a ReAct agent, with long-term and visual memory.
Automatic surface-defect detection on 1.5 cm semiconductor parts for HBM, in production at a Korean manufacturer.
LLM pipelines that turn trading and storage contracts into structured data: option value per contract, and what changed between contract versions.
From a Swiss address to the rules that apply, the legal building envelope, and an Archicad-ready 3D model.
Live demos at UN summits, a bank, schools and a TV studio, and two user studies with 123 participants.
Can GAN-generated images fix extreme class imbalance in sewer inspection? An honest negative result, and a labelling problem found on the way.
