Ziyue (Alvin) Liu
CV

About

I work on making LLM training cheaper, faster, and more reliable.

I am a Ph.D. student in Computer Science at the University of California, Santa Barbara, advised by Prof. Zheng Zhang. Before my Ph.D., I received an M.A. in Statistics from UCSB and a B.S. in Statistics from Nankai University.

My research covers the training stack end to end, particularly for pre-training: low-rank model architectures that reduce compute, Muon-style optimizers that train faster, and scalable, fault-tolerant training systems designed for clusters of 100k+ GPUs.

I also work on AI for EDA and data center thermal management, building ML surrogate models that speed up thermal and CFD simulation, from 3D-IC chips to data center digital twins.

Most recently, I was a Software Engineer Intern at Google Cloud AI. Before that, I was a visiting student at Argonne National Laboratory and a research intern at Cadence Design Systems.

I am on the job market for full-time industry research and engineering roles starting in 2027. Feel free to reach out at ziyueliu@ucsb.edu.

News

Scroll for earlier updates
  • ReCoVer was accepted to NeurIPS 2026.

  • Muon² was accepted to EMNLP 2026 as an oral presentation.

  • MuonQ was accepted to COLM 2026.

  • Started my internship at Google Cloud AI, on MoE adapter tuning and serving.

  • BOOST was accepted to MLSys 2026.

  • DeepOHeat-v1 was published in IEEE TCPMT.

  • LaX was accepted to NeurIPS 2025.

  • CoLA (oral) and QuZO were accepted to EMNLP 2025.

  • Visiting student at Argonne National Laboratory, on fault-tolerant LLM pre-training.

  • SepONet was published in TMLR.

  • CoMERA was accepted to NeurIPS 2024.

  • Third research internship at Cadence, on data center thermal and CFD modeling.

  • Started my Ph.D. in Computer Science at UCSB.

  • Back at Cadence as a research intern, on automotive aerodynamic simulation.

  • Received my M.A. in Statistics from UCSB.

  • DeepOHeat was accepted to DAC 2023.

  • TT-PINN was accepted to the ICML 2022 HAET workshop.

  • First research internship at Cadence, on 3D-IC thermal simulation.

  • Joined Prof. Zheng Zhang's group at UCSB.

  • Came to UCSB to start my M.A. in Statistics.

Research

My main research is on efficient LLM training, across three layers of the stack. My second line of work is AI for EDA and data center thermal management, built up through papers and three research internships at Cadence.

LLM training stack

Architecture

Low-rank and tensor-compressed models that cut the compute and memory cost of training while preserving quality.

Full-size layer
dWσdff
h = σ(Wx)
CoLA layer
dAσrBdff
h = B σ(Ax)

Optimizer

Muon-based optimizers that further speed up convergence, reduce orthogonalization cost, and improve quality.

GPU-hours to reach a loss of 2.36 on LLaMA-1B
Muon, 5 steps1,041 Muon², 3 steps816−22% Muon², 5 steps797−23% 03006009001,200

System

Scalable training for low-rank models, and fault-tolerant training that avoids system stalling on frequent restarts at O(100k+) scale.

AI for EDA / Data center

Chip

Physics-driven operator learning for 3D-IC thermal design that speeds up thermal simulation by orders of magnitude, handles multi-scale and multi-die designs, and enables fast, trustworthy thermal optimization.

Six unseen power maps; for each, the 3D temperature field from the Celsius 3D solver and from DeepOHeat look the same. Three unseen power maps; for each, the 3D temperature field from the Celsius 3D solver and from DeepOHeat look the same.

Data center

AI surrogates for data center digital twins, trained on data from physics-based solvers and integrated with Cadence's Reality Digital Twin Platform to predict airflow and server temperatures in near real time.

Methods

Faster and lighter physics-informed networks and neural operators, and CFD surrogates that scale to large automotive aerodynamic simulations.

Selected Publications

Full list on Google Scholar

First or co-first author* Equal contribution

Experience

Industry and research

  • Jun 2026 – Sep 2026

    Software Engineer Intern

    Google Cloud AI · Sunnyvale, CA

    Efficient MoE adapter tuning and multi-tenant serving.

  • Apr 2025 – Sep 2025

    Visiting Student

    Argonne National Laboratory · Lemont, IL

    Fault-tolerant LLM pre-training without checkpoint restarts, published as ReCoVer at NeurIPS 2026.

  • Jun 2022 – Sep 2024 · 3 terms

    Research Intern

    Cadence Design Systems · Austin, TX

    • 2024

      Data center thermal and CFD

      Simulation-trained AI surrogates for real-world data center thermal and CFD simulations, integrated with Cadence's Reality Digital Twin Platform to predict airflow and server temperatures in near real time.

    • 2023

      Automotive aerodynamics

      Data-driven models of large-scale automotive aerodynamic simulations.

    • 2022

      3D-IC thermal

      Physics-informed operator learning for 3D-IC thermal simulation, published as DeepOHeat at DAC 2023.

Education

  • 2023 – 2027 (expected)

    Ph.D. in Computer Science

    University of California, Santa Barbara

    Advisor: Prof. Zheng Zhang

  • 2021 – 2023

    M.A. in Statistics

    University of California, Santa Barbara

  • 2016 – 2020

    B.S. in Statistics

    Nankai University, School of Mathematical Sciences

Teaching

Teaching assistant at the University of California, Santa Barbara.

  • CMPSC 130AData Structures and AlgorithmsWinter 2025, Winter 2026
  • CMPSC 165BMachine LearningWinter 2024
  • PSTAT 231Statistical Machine LearningFall 2022, Winter 2023
  • PSTAT 120BProbability and StatisticsFall 2021, Spring 2022
  • PSTAT 5LSStatistics for Life SciencesWinter 2022