Taeyun Ha

I am a first-year PhD student at Seoul National University, advised by Prof. Hanbyul Joo. My research focuses on robotics and 3D computer vision.

I received my B.S. in Electrical and Computer Engineering, summa cum laude, from Seoul National University.

Google Scholar/ GitHub/ LinkedIn/ CV/ taeyun012@gmail.com

Research

I work on human-to-robot learning and dexterous manipulation, with an interest in making robot skills generalize beyond a single hand or environment.
Current interests
- Scaling human-to-robot learning: using human demonstrations and data to teach robot hands.
- Dexterous manipulation: learning robust grasping and interaction with diverse objects.
- Real-to-sim-to-real: connecting real-world data and simulation

Publications

AutoDex: An Automated Real-World System for Dexterous Grasping Data Collection

Mingi Choi, Gunhee Kim, Jisoo Kim, Taeksoo Kim, Taeyun Ha, Jongbin Lim, Hanbyul Joo

CoRL 2026

Project / Paper

An automated real-world data collection system for physically validated dexterous grasp trials using multi-view perception, robotic execution, outcome labeling, and autonomous reset.

HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments

Jongbin Lim^*, Taeyun Ha^*, Seongho Cha, Kanghyeon Cho, Mingi Choi, Subin Jeon, Jisoo Kim, Byungjun Kim, Hanbyul Joo

arXiv preprint, 2026

Project / arXiv / Dataset / Code

A multimodal dataset of high-fidelity dexterous grasping sequences featuring both human and robotic hands.

Projects

VR E-scooter Simulator

Woongchan Cho, Junhyeok Yang, Taeyun Ha

Project page

Developed an E-scooter simulator integrated with a Unity-based VR environment for a realistic riding experience.

Gold prize, 2024 SNU Creative Design Fair