avatar

Liangyu Zhao

赵亮雨

🏢 Member of Technical Staff · xAI xAI
🎓 Ph.D. Student · University of Washington
[myfirstname]@cs.washington.edu


About Me

I am a Member of Technical Staff at xAI and a Ph.D. candidate in the Paul G. Allen School of Computer Science & Engineering at the University of Washington, advised by Prof. Arvind Krishnamurthy. I also completed my undergraduate studies at the University of Washington, earning a B.S. in Computer Science and a B.S. in Applied & Computational Mathematical Sciences (Discrete Math and Algorithms).

My CV can be found here.

Research Interests

Machine learning systems, distributed systems, collective communications; broadly speaking, I am interested in applying mathematical techniques to design and build efficient, scalable computer systems.

News

Publications

  1. ForestColl: Throughput-Optimal Collective Communications on Heterogeneous Network Fabrics
    Liangyu Zhao, Saeed Maleki, Yuanhong Wang, Zezhou Wang, Ziyue Yang, Hossein Pourreza, Arvind Krishnamurthy
    USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2026.
    arXiv PDF Code Slides

  2. FAST: An Efficient Scheduler for All-to-All GPU Communication
    Yiran Lei, Dongjoo Lee, Liangyu Zhao, Daniar Kurniawan, Chanmyeong Kim, Heetaek Jeong, Changsu Kim, Hyeonseong Choi, Liangcheng Yu, Arvind Krishnamurthy, Justine Sherry, Eriko Nurvitadhi
    USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2026.
    arXiv PDF

  3. NanoFlow: Towards Optimal Large Language Model Serving Throughput
    Kan Zhu, Yufei Gao, Yilong Zhao, Liangyu Zhao, Gefei Zuo, Yile Gu, Dedong Xie, Tian Tang, Qinyu Xu, Zihao Ye, Keisuke Kamahori, Chien-Yu Lin, Ziren Wang, Stephanie Wang, Arvind Krishnamurthy, Baris Kasikci
    USENIX Symposium on Operating Systems Design and Implementation (OSDI), 2025.
    PDF Code

  4. Efficient Direct-Connect Topologies for Collective Communications
    Liangyu Zhao, Siddharth Pal, Tapan Chugh, Weiyang Wang, Jason Fantl, Prithwish Basu, Joud Khoury, Arvind Krishnamurthy
    USENIX Symposium on Networked Systems Design and Implementation (NSDI), 2025.
    PDF Talk Slides

  5. Rethinking Machine Learning Collective Communication as a Multi-Commodity Flow Problem
    Xuting Liu, Behnaz Arzani, Siva Kesava Reddy Kakarla, Liangyu Zhao, Vincent Liu, Miguel Castro, Srikanth Kandula, Luke Marshall
    ACM Special Interest Group on Data Communication (SIGCOMM), 2024.
    PDF Code

  6. Efficient all-to-all Collective Communication Schedules for Direct-connect Topologies
    Prithwish Basu, Liangyu Zhao, Jason Fantl, Siddharth Pal, Arvind Krishnamurthy, Joud Khoury
    International Symposium on High-Performance Parallel and Distributed Computing (HPDC), 2024.
    PDF

  7. AutoLRS: Automatic Learning-Rate Schedule by Bayesian Optimization on the Fly
    Yuchen Jin, Tianyi Zhou, Liangyu Zhao, Yibo Zhu, Chuanxiong Guo, Marco Canini, Arvind Krishnamurthy
    International Conference on Learning Representations (ICLR), 2021.
    PDF Code

  8. Nexus: A GPU Cluster Engine for Accelerating DNN-Based Video Analysis
    Haichen Shen, Lequn Chen, Yuchen Jin, Liangyu Zhao, Bingyu Kong, Matthai Philipose, Arvind Krishnamurthy, Ravi Sundaram
    ACM Symposium on Operating Systems Principles (SOSP), 2019.
    PDF Code

Recorded Talks

  1. Efficient Direct-Connect Topologies for Collective Communications
    USENIX NSDI
    Philadelphia, April 2025
    Video Slides

  2. Efficient Direct-Connect Topologies for Collective Communications
    FOCI Annual Symposium
    University of Washington, October 2023
    Video Slides


Powered by Jekyll and Minimal Light theme.