Research Engineer @ SMART Computer Vision · Robot Perception · Embodied AI Building intelligent systems that perceive, reconstruct, understand, and interact with the physical world.
Prior to joining SMART, I worked as a Research Associate at Nanyang Technological University (NTU), where I focused on computer vision, autonomous systems, and embedded AI under Professor Siew-Kei Lam. I also earned a Master of Engineering in Computer Science from NTU, where my dissertation addressed multi-camera tracking for smart urban mobility.
Interested in doing research in Computer Vision, Robot Perception, or Embodied AI?
I welcome motivated students to reach out about research opportunities. vmdo@mit.edu ·
LinkedIn ·
CV
Research
My research focuses on Computer Vision, Robot Perception, and Embodied AI. I am interested in building intelligent systems that can perceive and reconstruct the physical world, understand human-object interactions, and translate this understanding into reliable robot actions.
My current research spans visual perception and grounding, 3D geometric perception, interaction understanding and robot learning, industrial visual perception and inspection, and industrial simulation and benchmarking.
Featured Research Directions
Visual Perception & Grounding
Visual grounding, segmentation, and interaction-aware perception.
3D Geometric Perception
Depth, 6D pose, calibration, localization, and 3D reconstruction.
Third Prize — DUT Student Research Competition, 2016
Thesis: Robotic Arm Design for People with Arm Disabilities Using Head Gestures and Eyebrow Movements
Selected Previous Projects
Autonomous Drone in Maze
CCDS, NTU, Singapore · 2023–2024
Developed an autonomous drone system for real-time sensing, mapping, planning, and navigation in maze environments, with perception and navigation modules optimized for embedded computing on Jetson.
Developed edge-based visual analytics for infrastructure-to-vehicle communication, using real-time vehicle tracking and crowd counting from infrastructure cameras to support autonomous vehicle navigation near public transportation facilities.
Developed lightweight computer vision and deep learning models for real-time vehicle detection and tracking to reduce traffic-light waiting times, with continuous deployment on resource-constrained embedded platforms.
Improved visual SLAM robustness through dynamic outlier removal and semantic loop closure for RGB-D and stereo systems, focusing on pose estimation, mapping, and deployment on embedded platforms.
Developed lightweight computer vision and deep learning models for real-time bus detection and tracking on low-cost embedded platforms, supporting operation across changing weather, illumination, and traffic conditions.
