CV
After nearly five years building computer vision and mapping systems for autonomous vehicles at NVIDIA, I recently left to pursue the startup path. I’m now CTO / Founding Engineer at Athleo.ai, an AI-powered sports video storytelling platform, while developing the TransFOMO app as a pet project on the side. I’m always happy to connect — and remain open to interesting robotics / computer vision opportunities.
Education
- M.S. in Robotics System Development, Carnegie Mellon University, School of Computer Science, 2021
- B.S. in Electrical Engineering, National Taiwan University, 2019
Skills
- Programming Languages: Python, C/C++, ROS, SQL
- Tools: OpenCV, Pytorch, Bazel, Docker
- Computing Environments: Linux, Windows, MacOS
Work experience
- Apr 2026-Present: CTO / Founding Engineer
- Athleo.ai, an AI-powered sports video storytelling platform that turns raw sports footage into professional-quality highlights.
- Jun 2023-Apr 2026: Senior Computer Vision Engineer (Auto-labeling team)
- NVIDIA, Redmond, WA
- Developed a 4D multi-object tracking autolabeling pipeline for autonomous vehicles (AV) by integrating perception results from LiDAR or camera, with production-level ML model evaluation and release/deployment.
- Constructed the cuboid data ingestion pipeline with sensor fusion and motion compensation for LiDAR.
- Built visualization tools for overlaying cuboids on LiDAR/camera and Redash dashboards for KPI tracking.
- Person-in-charge for pipeline bring-up on a new sensor suite.
- Jul 2021-Jun 2023: Computer Vision Engineer (mapping team)
- NVIDIA, Redmond, WA
- Optimized the in-car real-time mapping software, reducing 31% of memory bandwidth without losing useful information downstream while still meeting product requirements.
- Implemented lane graph generation for an offline AV HD mapping system with complex BFS & DFS algorithms.
- Contributed to system design, software optimization, and testing to improve mapping accuracy and efficiency.
- May-Aug 2020: Computer Vision Intern (mapping team)
- NVIDIA, Redmond, WA
- Designed and implemented the wait-line detection pipeline for NVIDIA’s in-car mapping system.
- Feb-Jun 2019: Machine Learning Intern
- Cardinal Blue Software, Taipei, Taiwan
- Trained a semantic segmentation model with Mobilenetv2 structure that cut out people from an image and deployed it to the PicCollage mobile app.
- Achieved 83% IOU with model size <10mb, 6% better than the official released version.
- Jul-Aug 2018: Summer Intern
- Coretronic Corporation, Hsinchu, Taiwan
- Constructed independently a monocular visual odometry system aimed for drones.
- Implemented ORB feature extraction with OpenCV and local pose optimization with G2O in C++.
Selected Projects
- AutoValet - Autonomous Parking System in Indoor Environment, Carnegie Mellon University, Sep 2019-May 2021
- Made the robot Husky perform SLAM in an unknown indoor parking garage and combine geometrically constrained exploration with navigation to find an empty parking spot, and park itself accurately.
- Contributed in building Gazebo simulation, visual lane detection, parking algorithms, robot exploration, and SLAM.
- Visual Search with Semantic Segmentation on ORB SLAM, National Taiwan University, Sep 2018-Jun 2019
- Combined ORB SLAM and semantic segmentation on each frame into a real-time system that builds semantic point cloud
- Designed an effective algorithm to localize target objects in 3D maps.
- Real-time 3D Reconstruction Using Visual SLAM, National Taiwan University, Jan-Jun 2018
- Used Zed stereo camera and Kinect 2 RGB-D camera to implement mapping and localization.
- Developed Real-Time Appearance-Based Mapping (RTAB-Map) with ROS.
- Robots Gain Social Intelligence Through Reinforcement Learning, National Taiwan University, Jun 2018
- Trained the humanoid robot Pepper to learn basic social skills through Proximal Policy Optimization in deep rein- forcement learning, including understanding in what situation it should greet, say bye-bye, shake hands, or simply do nothing.
- Service Robot for Deaf and Speech-impaired People, National Taiwan University, Oct-Dec 2017
- Trained a convolutional neuron network to recognize sign language videos.
- Implemented the system to a service humanoid robot to assist deaf and speech-impaired people.
Leadership & Awards
- STMicroelectronics Sponsor Prize, 2019 MakeNTU LAzITech Hackathon, Mar 2019
- Academic Coordinator, Electrical Engineering Student Association, National Taiwan University, Sep 2017-Jun 2018
- General Coordinator, 2018 MakeNTU AInsight Hackathon, Mar 2018
- Led a team over 70 members to organize this 5-day hackathon about AI for undergraduates all over Taiwan.
- Touched base with Ministry of Education, Ministry of Science and Technology, Industrial Technology Research Insti- tute, and plenty of companies for partnership and sponsorship.
- Check our website here and Facebook fanpage here.
- Bronze Medal & HUNSTEM Special Prize, 2014 International Sustainable World Olympaid - I-SWEEEP, May 2014