Zhaoheng (Billy) Li

CS Ph.D. Student @ University of Illinois Urbana-Champaign

prof_pic.jpg

Thomas M. Siebel Center, Room 2119

201 N. Goodwin Ave.

Urbana, IL 61801

I am a fifth-year CS PhD student at the University of Illinois at Urbana-Champaign (UIUC) advised by Prof. Yongjoo Park. I am a member of CreateLab and DAIS. My research interests are Systems for AI and ML, Interactive Data Analytics, Computational Notebooks, and Vector Search & Databases.

My research targets building systems for emerging AI and Data Science applications with robustness and scalability guarantees; I am currently working on data infrastuctures for Agentic Data Science Frameworks. My research has been generously funded by the NCSA for the Fall 2023 semester.

During my undergraduate study at UIUC, I was fortunate to get early exposure to research, starting with a fun side project of DOTA2 match outcome prediction in the Esports Analytics Lab. I then proceeded with involved works: Deep Steerable Graph Generation with Prof. Carl Yang, and my bachelor’s thesis “REFORM: Fast and Adaptive Solution for Subteam Replacement” advised by Prof. Hanghang Tong.

I dedicate my summers to internships; My previous internships and projects include:

  • Summer 2025: System Infrastructure Lab @ ByteDance, Cloud-Native Vector Indexes
  • Summer 2024: System Infrastructure Lab @ ByteDance, Filtered Vector Search Indexing
  • Summer 2023: Google BigQuery @ Google, Efficient GROUP BY for Structs
  • Summer 2022: Google S2Infra @ Google, SQL Profiling with BPF
  • Summer 2020: Google Local Services @ Google, Ads Ranking Algorithms
  • Summer 2019: Google ContentAds @ Google, Monitoring Pipeline for Ad Requests

News

Nov 15, 2025 Our research paper, “Chipmink: Efficient Delta Identification for Massive Object Graph”, has been accepted to VLDB 2026.
Nov 04, 2025 I gave a talk at JupyterCon 2025 titled “World’s First Undoable Notebook” for my ongoing research project, Kishu. → Talk
Oct 20, 2025 Our research paper, “MojoFrame: Dataframe Library in Mojo Language”, has been accepted to ICDE 2026.
Oct 16, 2025 Our research paper, “QStore: Quantization-Aware Compressed Model Storage”, has been accepted to VLDB 2026.
Jun 30, 2025 Our demo, “Demo of Kishu: Time-Traveling for Computational Notebooks”, received the Best Demo Award at SIGMOD 2025!

Selected Publications

  1. Cloud-Native Vector Search: A Comprehensive Performance Analysis
    Zhaoheng Li, Wei Ding, Silu Huang, Zikang Wang, Yuanjin Lin, Ke Wu, Yongjoo Park, and Jianjun Chen
    2026
  2. Chipmink: Efficient Delta Identification for Massive Object Graph
    Supawit Chockchowwat, Sumay Thakurdesai, Zhaoheng Li, Matthew Krafczyk, and Yongjoo Park
    Proceedings of the VLDB Endowment, 2026
  3. MojoFrame: Dataframe Library in Mojo Language
    Shengya Huang, Zhaoheng Li, Derek Werner, and Yongjoo Park
    In 42nd IEEE International Conference on Data Engineering, ICDE 2026, 2026
  4. QStore: Quantization-Aware Compressed Model Storage
    Raunak Shah, Zhaoheng Li, and Yongjoo Park
    Proceedings of the VLDB Endowment, 2026
  5. SIEVE: Effective Filtered Vector Search with Collection of Indexes
    Zhaoheng Li, Silu Huang, Wei Ding, Yongjoo Park, and Jianjun Chen
    Proceedings of the VLDB Endowment, 2025
  6. Kishu: Time-Traveling for Computational Notebooks
    Zhaoheng Li, Supawit Chockchowwat, Ribhav Sahu, Areet Sheth, and Yongjoo Park
    Proceedings of the VLDB Endowment, 2024
  7. demo
    Demonstration of Kishu: Time-Traveling for Computational Notebooks (SIGMOD 2025 Best Demo)
    Zhaoheng Li, Supawit Chockchowwat, Hanxi Fang, and Yongjoo Park
    In Companion of the 2025 International Conference on Management of Data, 2025
  8. ElasticNotebook: Enabling Live Migration for Computational Notebooks
    Zhaoheng Li, Pranav Gor, Rahul Prabhu, Hui Yu, Yuzhou Mao, and Yongjoo Park
    Proceedings of the VLDB Endowment, 2023
  9. S/C: Speeding up Data Materialization with Bounded Memory
    Zhaoheng Li, Xinyu Pi, and Yongjoo Park
    In 39th IEEE International Conference on Data Engineering, ICDE 2023, 2023

Research Projects

Here are some projects I built that I conveniently have demos for:

Kishu: Versioned and Undoable Notebook System

Motivated by my frustrations with using Colab for research when I was an undergrad, Kishu is a Jupyter-based notebook system where users can undo cell executions to 'un-drop' dataframe columns, restore overwritten models, etc.

Hobbies

Badminton, Biking, Competitive DOTA2 (proud member of the UIUC DOTA2 varsity team during the 2017-2018 academic year), Rhythm Games.