Lee Yeong Khang

Work

Senior R&D Engineer, ViTrox Technologies
2018 - Present, Bayan Lepas, Penang

  • Lead development of an AI training platform covering image labeling, model training, deployment, and production support.
  • Support computer vision engineers, technical support, sales engineers, and customers using the training platform.
  • Maintain and deploy deep learning models including ResNet, Faster R-CNN, YoloX, YoloV8, SegFormer, UNet, Vision Transformer, PatchCore, and PaDiM.
  • Built models for 360-degree OCR orientation detection, anomaly detection, and segmentation.
  • Set up shared GPU infrastructure, including a 32-GPU cluster, Multi-Instance GPU allocation, and an internal Jupyter Notebook environment.
  • Worked on AOI post-verification systems using deep learning for high-mix, high-volume production data.
  • Led an AI-assisted code review effort using Gemini and GitLab integration.
  • Collaborated with researchers including Dr. Chan Chee Seng on OCR model research and publication work.

R&D Engineer, ViTrox Technologies
2015 - 2018, Bayan Lepas, Penang

  • Started a data analytics and processing platform using Python, pandas, Flask, and multiple data sources including PostgreSQL, MySQL, MSSQL, Oracle, Excel, and Google Sheets.
  • Designed workflow systems for monitoring, logic automation, and anomaly notification.
  • Established Kubernetes and Docker infrastructure for scalable analytics workloads and lightweight edge deployment.
  • Introduced Git source control, GitLab, and CI/CD automation for engineering teams.
  • Piloted deep learning classification models on AOI inspection data and taught internal computer vision tutorials.

Research

My research background is in pattern recognition, machine learning, biometrics, long-tailed visual recognition, OCR, anomaly detection, and face recognition on tensor manifolds.

Selected publications:

  • CONSULT: Contrastive Self-Supervised Learning for Few-shot Tumor Detection
    Sin Chee Chin, Xuan Zhang, Lee Yeong Khang, Wenming Yang. arXiv, 2024.
  • Rethinking Long-Tailed Visual Recognition with Dynamic Probability Smoothing and Frequency Weighted Focusing
    Yeong Khang Lee, Cheng Yaw Low, Andrew B.J. Teoh. IEEE ICIP, 2023.
  • ICDAR 2021 Competition on Integrated Circuit Text Spotting and Aesthetic Assessment
    Chun Chet Ng, Akmalul Khairi Bin Nazaruddin, Yeong Khang Lee. ICDAR, 2021.
  • Tensor Kernel Supervised Dictionary Learning for Face Recognition
    Yeong Khang Lee, Cheng Yaw Low, Andrew B.J. Teoh. APSIPA ASC, 2015.
  • Joint Kernel Collaborative Representation on Tensor Manifold for Face Recognition
    Yeong Khang Lee, Andrew B.J. Teoh, Kar-Ann Toh. ICASSP, 2014.
  • A Study on Distance Measures of Tensor Manifold for Face Recognition
    Yeong Khang Lee, Andrew B.J. Teoh. ICEIC, 2014.

Education

Yonsei University
Master in Electronic and Electrical Engineering, 2013 - 2015
Multimedia Security Lab, supervised by Andrew Teoh Beng Jin. Thesis: Face Recognition on the Tensor Manifold.

Multimedia University
Bachelor of Information Technology (Honours), Security Technology, 2008 - 2012
Relevant project: UDP encapsulation on IPsec for heterogeneous networks.