About me

Postdoctoral Researcher in the Laboratory of Artificial Intelligence and Robotics (LAIR) at Sungkyunkwan University (SKKU).
My research focuses on Reinforcement Learning (RL) for robotics.

Work Experience

Epitone, Inc. (San Diego, USA) — Sr. Software Engineer / AI Researcher (Nov 2023 - Apr 2025)
  • Developed lightweight CNN for pupil detection
  • Designed Bayesian DNN for confidence estimation
  • Converted models to TensorRT for < 1ms inference
  • Integrated tracker with Kalman Filter and Optical Flow
  • Built and labeled proprietary pupil dataset
SoLiD (Gyeonggi-do, South Korea) — Industrial Technical Personnel (May 2010 - July 2012)
  • Simulated repeater automation and managed database systems

Education

  • Post-Doc in Department of Intelligent Robotics, Sungkyunkwan University (2025–present)
  • Ph.D. in Department of Electrical and Computer Engineering, Purdue University (2016–2023)
  • M.S. in Department of Electrical and Computer Engineering, Seoul National University (2013–2015)
  • B.S. in Department of Electrical and Computer Engineering, Seoul National University (2007–2013)

Projects

High-Speed Deployment of LSeg Image Encoder
  • Extracted and optimized the image encoder from the LSeg segmentation model for real-time semantic segmentation.
  • Converted to ONNX and optimized with TensorRT in C++ pipeline, significantly improving inference speed.
  • Tools: PyTorch, TensorRT, OpenCV, Python, C++
Unsupervised Object Tracking via Video Colorization
  • Implemented video colorization-based tracker using TensorFlow (based on Vondrick et al.).
  • Demonstrated unsupervised object tracking using color consistency in video.
  • Tools: TensorFlow, Python, OpenCV

Publications

  • [J1] J. Moon, D. Das, C. S. G. Lee.
    "A Multistage Framework With Mean Subspace Computation and Recursive Feedback for Online Unsupervised Domain Adaptation."
    IEEE Transactions on Image Processing, 2022.
    DOI: 10.1109/TIP.2022.3186537
  • [C1] J. Moon, D. Das, C. S. G. Lee.
    "Multi-step Online Unsupervised Domain Adaptation."
    ICASSP, 2020.
    DOI: 10.1109/ICASSP40776.2020.9052976
  • [C2] D. Das, J. Moon, C. S. G. Lee.
    "Few-shot Image Recognition with Manifolds."
    ISVC, 2020.
    DOI: 10.1007/978-3-030-64559-5_1

Skills

  • Programming Languages: Python, C++, C, Matlab, SQL, Excel VBA, R, LaTeX
  • Machine Learning & Data Science: PyTorch, TensorFlow, Scikit-Learn, Pandas
  • Computer Vision & Optimization: OpenCV, MediaPipe, Kalman Filter, Optical Flow, Bayesian Estimation
  • Software Engineering & Development: Docker, Git, CMake, Bash Scripting
  • Robotics & Embedded Systems: ROS, RealSense SDK

Awards and Honors

  • NSF Grant (IIS-1813935), co-author
  • NVIDIA Academic Hardware Grant 2018
  • Summer Research Grant, Purdue University

Service

  • Reviewer: IEEE TCSS, IEEE TIP, JIFS

Teaching

  • TA: Electronic Devices and Design Lab @ Purdue (2017–2020)

Hobbies

Calisthenics, Swimming, Badminton, Tennis, Squash, Table-tennis, Guitar