Resume

A public, web-safe overview of my engineering experience, education, publications, and patents.

Contact Information

Name Kevin Chen
Professional Title Spatial AI & On-Device ML Systems Engineer
Email kevchen.us@gmail.com
Location San Francisco Bay Area, CA
Website https://kuantingchen04.github.io

Experience

  • 2021 - present

    Palo Alto, CA

    Senior Computer Vision Engineer
    OPPO US Research Center
    3D mapping, robust localization, and real-time vision for XR and AI devices.
    • Tech lead for multi-camera and IMU VIO/SLAM, delivering real-time state estimation under strict latency and power constraints with cm-level ATE on long sequences.
    • Co-authored ActiveGAMER (CVPR 2025) and built its data preprocessing plus reproducible training and evaluation pipeline for Gaussian scene representations.
    • Trained and integrated a learned keypoint and descriptor frontend, including quantized on-device inference and regression evaluation for matching and tracking stability.
    • Built depth and TSDF reconstruction pipelines, open-vocabulary localization integrations, and system validation tooling for mapping quality and drift.
    • Profiled and optimized memory, CPU parallelism, and data flow while moving algorithms from research prototypes to mobile devices.
  • 2018 - 2021

    Palo Alto, CA

    Computer Vision Engineer
    OPPO US Research Center
    SLAM frontend and mobile computer vision systems.
    • Implemented feature tracking, matching, pose estimation, and camera-model support for monocular, stereo, and multi-camera AR pipelines.
    • Built plane and surface understanding modules from SLAM maps and depth images for AR interaction.
    • Optimized core computer vision components with ARM NEON and x86 SIMD for constrained mobile SoCs.
    • Developed a reusable C++ computer vision library shared across multiple product and research pipelines.
  • 2017 - 2017

    Ann Arbor, MI

    Research Assistant
    University of Michigan Ford Center for Autonomous Vehicles
    • Developed a multi-object tracking system combining RGB cameras, LiDAR, Kalman filtering, object detection, and re-identification.
  • 2014 - 2015

    Hsinchu, Taiwan

    Research Assistant
    National Tsing Hua University Vision Science Lab
    • Built a large-scale Structure-from-Motion and trajectory-alignment framework for dashcam localization and street-view change detection.

Summary

  • Senior computer vision engineer with 8+ years of experience building real-time spatial AI systems across SLAM/VIO, 3D mapping, learned visual features, and on-device ML. Strong C++ and Python background spanning research pipelines, evaluation infrastructure, and deployment on resource-constrained XR and AI devices.

Education

  • 2016 - 2018

    Ann Arbor, MI

    M.S.
    University of Michigan
    Electrical and Computer Engineering - Computer Vision and Artificial Intelligence
    • Developed a GAN-based light-field depth estimation project with faster processing and 90% data reduction.
  • 2011 - 2015

    Hsinchu, Taiwan

    B.S.
    National Tsing Hua University
    Electrical Engineering
    • Academic Achievement Awards, 2013-2015.

Skills

Languages: C, C++, Python, Matlab, Java, JavaScript
Spatial AI: SLAM, VIO, multi-view geometry, camera calibration, 3D reconstruction, TSDF, Gaussian representations
CV and ML: PyTorch, TensorFlow, OpenCV, Eigen, ROS, PCL, learned features, vision-language models
Systems: profiling, latency and memory optimization, ARM NEON, CPU and DSP parallelism, quantization, Android
Infrastructure: Git, Linux, Docker, Jenkins, AWS, REST APIs, SQL

Publications

Patents

2024: An Object Search System using 3D Language Field
2024: System and Method for Optical-Flow Prediction via Knowledge Transferring from Large Vision Models
2023: Exposure Control to Handle Tracking and Stereo Matching for Multi-Camera VIO/SLAM System
2022: An Indoor Wi-Fi Access Point Localization Method using VIO/SLAM
2020: Method and System for Adaptive Feature Detection for VSLAM Systems