Publications

* indicates equal contribution

2026

  1. ICML’26 (In Submission)
    HELIOS.png
    HELIOS: Heterogeneous Lightweight VLA Model Serving System
    Jongheon Jeong*, Hojoon Kim*, Rokhee Lee, Yeonhong Park, Young H. Oh, and Jae W. Lee
    2026
    In submission
  2. MLSys’26
    agenticcache.jpg
    AgenticCache: Cache-Driven Asynchronous Planning for Embodied AI Agents
    Hojoon Kim, Yuheng Wu, and Thierry Tambe
    2026
  3. ICML’26 (In Submission)
    queso.jpg
    QUESO: Storage-Assisted Quantization Error Compensation for On-Device LLM Inference
    Seong Hoon Seo, Donghyun Lee, Geonha Lee, Hojoon Kim, Yeonhong Park, and Jae W. Lee
    2026
    In submission

2025

  1. ICML’25 Spotlight
    flashtp.jpg
    FlashTP: Fused, Sparsity-Aware Tensor Product for Machine Learning Interatomic Potentials
    Seung Yul Lee, Hojoon Kim, Yutack Park, Dawoon Jeong, Seungwu Han, Yeonhong Park, and Jae W. Lee
    2025
    Spotlight (313/12,107, 2.6%)
    PDF Code
  2. OSDI’25
    decdec.jpg
    DecDEC: A Systems Approach to Advancing Low-Bit LLM Quantization
    Yeonhong Park*, Jake Hyun*, Hojoon Kim, and Jae W. Lee
    2025
    Acceptance Rate: 52/327 = 15.9%
    PDF Code