Hi, welcome!

I am a fifth-year Ph.D. student in Computer Science at the University of California, Berkeley. I am advised by Professor Alvin Cheung. I am a part of EPIC Data Lab and Sky Lab (formerly RISE Lab). I am interested in Efficient Video Inference Systems.

I graduated with a Combined Bachelor's/Master's program in Computer Science and Engineering from the University of Washington in 2021. I worked in the UW Interactive Data Lab as a research assistant, mentored by Professor Jeffrey Heer. I interned at Google as a student researcher and DocuSign and OctoAI as a software engineering intern. Aside from academic works, I was a volunteer officer of Thai Students Association, coordinating social events and meetups for Thai students in the University of Washington.


Selected Projects

Tetris Efficient and High-Fidelity Video Object Tracking.  Site  Demo
Spatialyze A Geospatial Video Analytic System.  Site  DOI
Vega-Label A fast labeling algorithm to layout labels annotating data points in charts.  GitHub
HCI Study Guide A study guide for UC Berkeley EECS HCI prelim exam.  Site
Arquero-SQL In progress: SQL back-end support for Arquero queries.  GitHub
Pokemon Go Forecast A web application that predicts weather in Pokemon Go for players to maximize game rewards.  Site  GitHub

Publications

Tetris: Tile-level Sampling for Efficient and High-Fidelity Video Object Tracking

2026.  arXiv  Site  Demo  GitHub

Spatialyze: A Geospatial Video Analytics System with Spatial-Aware Optimizations

VLDB 2024.  arXiv  VLDB  DOI  Site  GitHub

Towards Auto-Generated Data Systems

Shadaj Laddad  ·  Xiaoxuan Liu  ·  Chenglong Wang  ·  Cong Yan
VLDB 2023.  DOI  Paper

Design Study for a Geospatial-Video Data Analysis Query Language

2022.

Efficient Distributed Data Loading for Large-Scale Machine Learning Model Training with Parax

2021.  Paper

Legible Label Layout for Data Visualization, Algorithm and Integration into Vega-Lite

Master's Thesis 2021.  arXiv

Community Cellular Networks Coverage Visualizer

2021.  Paper  GitHub

Fast and Flexible Overlap Detection for Chart Labeling with Occupancy Bitmap

IEEE VIS 2020.  DOI  VIS  Video

Pleiades: Interactive Composing Tools for Vega-Lite Charts

2019.  Paper  GitHub

Work Experience

  • 2026 - present

    ML Software Engineering Intern

    Meta (Meta Recommendation Systems Knowledge Team)

    (On-going) Design and implement efficient Video-LLM (with Dr. Litao Hu).

  • 2021 - present

    Graduate Student Researcher

    Sky Lab @ UC Berkeley

    Contirbuting to Apperception, a query language and a database management system for geo-spatial video data. I am working on designing the programming interface of Apperception DSL. I am also working on video query optimization, using geo-spatial metadata to reduce executions of expensive ML operations.

  • 2025

    Student Researcher (Intern)

    Google (SystemResearch@Google)

    Researched Gemini’s video understanding evaluation through auto-generated video-QAs. Develop scalable question auto-generation methods and a correctness verifier for the generated questions (with Dr. Samira Khan).

  • 2021

    Software Engineering Intern

    OctoAI

    Integrate the deep-learning models visualizers designed from previous year into optimizer tool (Octomizer) that optimizes deep-learning models compiled by TVM.

  • 2019 - 2021

    Graduate and Undergraduate Teaching Assistant

    Paul G. Allen School of Computer Science and Engineering @ UW

    Teaching CSE331 (Software Design and Implementation) and CSE442 (Data Visualization). Teaches supplemental materials in quiz section. Grade students' assignments. And, hold office hours for answering questions related to the class.

  • 2018 - 2021

    Graduate and Undergraduate Research Assistant

    Interactive Data Lab @ UW

    Contributing to Vega and Vega-Lite project; both are toolkits for creating visualizations. In Vega, I added support for automated labeling data points in charts. In Vega-Lite, I added support for visualizing data with error bar or error band chart.

  • 2020

    Software Engineering Intern

    OctoAI

    Design and create visualization for deep-learning models and their performance using D3/TypeScript. The visualization is designed to be a part of optimizer tool (Octomizer) that optimizes deep-learning models compiled by TVM and measures their performances.

  • 2019

    Software Engineering Intern

    DocuSign

    Contributing to designing and developing a data lake platform that extracts, sanitizes, and stores users'/internal usage data into Amazon Web Service S3. Then, Data Scientist can query the data to perform analytics.

  • 2017

    Undergraduate Research Assistant

    Department of Computer Science @ Kasetsart University

    Categorize students' theses in the university's library based on language features of their abstracts.


Mentorship

Current

  • Alena Hou Chao

    2025 Working on the Tetris Project EECS BS at UC Berkeley

  • Charlie Si

    2025 Working on the Tetris Project CS BS at UC Berkeley

Former

  • Yongming Ge

    2021 - 2024 Worked on the Sptialyze Project. Apple

  • Yousef Helal

    2021 - 2024 Worked on the Sptialyze Project.