[CV | LinkedIn | GitHub | Google Scholar]

I’m a software engineer at Google working on systems for LLM performance tuning and eval. I also contribute to Gemini post training and XLA compiler optimizations.

I’m broadly interested in systems. My graduate school research focused on serverless computing problems such as orchestrating large-scale serverless applications and speeding up lightweight virtual machines. Previously, I worked on security and privacy, first for Android and later for IoT. Even earlier I wrote Linux kernel drivers for storage servers when I occasionally debugged by looking at how the LEDs on the motherboard lit up.

Work Experience

Google, Senior Software Engineer

  • GenAI Systems, Google Deepmind (2025-)
  • TPU Performance (2024-2025)
    • Systems for fleet-scale ML performance tuning and eval
    • XLA compiler auto-optimizations
    • LLM post training

Nimble Storage, Member of Technical Staff, 2015 - 2017

  • Linux kernel driver developer for storage servers

Education

Princeton University, 2017 - 2023

Duke University, 2011 - 2015

  • B.S.E in Electrical and Computer Engineering, minor in mathematics

Papers

  • Doing More with Less: Orchestrating Serverless Applications without an Orchestrator
    David H. Liu, Amit Levy, Shadi Noghabi, Sebastian Burckhardt
    Proc. 20th Symposium on Networked Systems Design and Implementation
    (NSDI ‘23) Boston, MA. April 2023.
    [paper] [slides] [code] [video]

  • How Low Can You Go? Practical cold-start performance limits in FaaS
    Yue Tan, David H. Liu, Nanqinqin Li, Amit Levy
    ArXiv Technical Report:2109.13319, Sept. 2021
    [paper]

  • Pyronia: Intra-Process Access Control for IoT Applications
    Marcela S. Melara, David H. Liu, Michael J. Freedman
    ArXiv Technical Report:1903.01950, March 2019
    [paper] [poster]

  • Sandtrap: Tracking information flows on demand with parallel permissions
    Ali Razeen, David H. Liu, Alvin R. Lebeck, Alexander Meijer, Valentin Pistol, Landon P. Cox
    Proc. 16th Annual International Conference on Mobile Systems, Applications, and Services
    (MobiSys ‘18) Munich, Germany, June 2018
    [paper]