Available for research & engineering work
Hi There
About
Sangyun Bae
2026.03 — 2026.09
Intern
Advanced Robot Control Lab, Center for Humanoid Robotics, KIST
Ongoing
2025.09 — 2026.02
Intern
ALLAI Inc.
Industry
2025.07 — 2025.08
NIPA NVIDIA-AI 전문가 과정
KOSA · Excellence Prize · Outstanding Final Project
Award
2025.05 — 2025.12
University Student Mentor
Incheon Ganghwa Office of Education
Mentoring
2021.03 — 2027.02
B.S. Information Convergence
Kwangwoon University (Expected)
Degree
NVIDIA DLI 교육과정
Certified Courses.
5 credentials
01
CV · Industrial Inspection
Computer Vision for Industrial Inspection
DLI-01
PDF ›
02
Synthetic Data for CV
Bootstrapping Computer Vision Models with Synthetic Data
DLI-02
PDF ›
03
OpenUSD · Stages & Prims
Learn OpenUSD — Stages, Prims and Attributes
DLI-03
PDF ›
04
Modern CUDA C++
Fundamentals of Accelerated Computing with Modern CUDA C++
DLI-04
PDF ›
05
CUDA Python
Fundamentals of Accelerated Computing with CUDA Python
DLI-05
PDF ›
Languages
Python
C / C++
TypeScript
CUDA
ML / DL
PyTorch
TensorRT
Robotics
ROS 1 / 2
Isaac Lab
MuJoCo
Jetson
Web / Data
Next.js
MongoDB
Work
Projects. What I can do.
Writing
글
My technical research, daily thoughts, paper notes, and more.
VAE 1편. 변분
VAE 1편 변분은 무엇인가
GAI
›
RTI (Random Torque injection)
Domain Randomization -> Torque Perturbation
Paper
›
PACE
sim-real gap 을 줄이기 위한 로봇의 동역학적 파라미터들을 탐색에 대한 알고리즘
파라미터 탐색 알고리즘은 CMA-Es 를 사용하고, 로봇의 config 에 들어가는 동역학 파라미터 값들을 찾는 것이 목표이다. (time delay, damping, 마찰계수 등등)
데이터 자체는 실제 데이터를 수집하고, Trajectory residual 이 가...
Paper
›
TRPO 에 대해
About TRPO (Trust Region Policy Optimization) which apply constraints to KL divergence called Trust Region
Paper
›
A3C 에 대해
DQN 에 이어 비동기 전략을 통한 학습전략을 사용한 A3C
Paper
›