Robot learning · Dexterous manipulation
Youngwoong Cho
I work on getting robot policies to transfer across embodiments — from one dexterous hand to another, and from human demonstrations to robots.
M.S. Robotics student at Georgia Tech in Prof. Danfei Xu's Robot Learning and Reasoning Lab. Previously a Research Engineer at RLWRLD, adapting robot foundation models to dexterous hands.
Interactive
One action interface, seven hands
Every hand gets the same target: a wrist pose and five fingertip positions. Per-hand inverse kinematics turns it into joint commands — the interface behind my cross-embodiment study. Pick a grasp, drag the slider, or orbit the scene.
Projects
One Action Interface, Seven Dexterous Hands
Does training on more kinds of robot hands let a policy control a hand it has never seen?
- Ongoing
- Simulation
- Cross-embodiment
- Teleoperation
What Makes Retargeted Demonstrations Learnable?
The same human demonstrations, retargeted two ways, train policies with very different success. Which properties of the retargeted actions explain it?
- Ongoing
- Real robot + simulation
- Imitation learning
- Humanoid
Adapting Robot Foundation Models to a 16-DoF Hand
Bringing large pretrained policies — GR00T N1.6, π0.5, and a 3D geometry-aware policy — to DexJoCo, a public 11-task dexterous-manipulation benchmark.
- Industry
- Simulation (MuJoCo)
- VLA fine-tuning
- Evaluation
Experience
- Aug 2026 – present
Georgia Tech — Robot Learning and Reasoning Lab (RL2)
Graduate Researcher, advised by Prof. Danfei Xu · Atlanta, GA
Cross-embodiment dexterous manipulation; learnability of retargeted demonstrations.
- Apr – Aug 2026
RLWRLD
Research Engineer · Seoul, Korea
Adapted robot foundation models (GR00T N1.6, π0.5, GAM) to dexterous hands; built the train/eval platform.
- Mar – Aug 2022
Naver Labs
Deep Learning Research Intern, Autonomous Driving · Seongnam, Korea
RL and imitation learning for 3D annotation; built a 3D annotation tool for collecting human demonstrations.
- May 2021 – Mar 2022
Seoul Robotics
Machine Learning Research Engineer Intern · Seoul, Korea
LiDAR perception: curb detection, multi-view annotation tooling, evaluation pipeline.
- May – Jul 2020
Seoul National University — Interactive & Network Robotics Lab
Research Intern, Prof. Dongjun Lee · Seoul, Korea
Cooperative manipulation on a physical multi-robot testbed.
Education
- Expected May 2028
Georgia Institute of Technology
M.S. in Robotics
- 2023
The Cooper Union
B.E. in Mechanical Engineering, Minor in Computer Science — Summa Cum Laude