Ziliang LAI

I am a Senior Software Engineer at Huawei’s Gauss Data Lab, building AI data infrastructure with a focus on unstructured data processing and Agents for data analytics. Our system is now powering serveral LLM data production lines in Huawei including those for training Xiaoyi (Huawei’s AI assistant). I hold a Ph.D. in Computer Science from CUHK, where my research centered on distributed databases.

Experiences

Senior Software Engineer @ Huawei Gauss Data Lab

Dec 2023 until now

Project: AI-centric Unstructured Data Processing Engine
Outcome: The system has been battle-tested in production with a cluster over 1000 Ascend NPUs (Huawei’s GPU); significantly reduced computational resource costs and time-to-train for internal LLMs up to 7x.
Technical Execution:
- Solution design: architected the core processing engine, featuring a user-friendly Python API, a high-performance C++ backend, and a serverless foundation
- Optimization: developed dynamic scheduling for heterogeneous hardware (CPU/Ascend NPU) and optimized offline batch LLM inference
- Ecosystem & Integration: integrated with Iceberg, Lance and Faiss for storage and indexing; implemented automatic pipeline orchestration with Agentic AI

Research Inten @Alibaba DAMO Academy

Sept 2021 - Jun 2022

Research: General, strongly consistent, and high-performance distributed transaction protocols.
Outcome (two papers):
- [ICDE 2023] Knock Out 2PC with Practicality Intact: A High-performance and General Distributed Transaction Protocol
- [SOCC 2024] Occam’s Razor for Distributed Protocols

Publications

Occam’s Razor for Distributed Protocols

Ziliang Lai, Fan Cui, Eric Lo, Wenchao Zhou, Feifei Li [paper]
Knock Out 2PC with Practicality Intact: a High-performance and General Distributed Transaction Protocol

Ziliang Lai, Hua Fan, Wenchao Zhou, Zhanfeng Ma, Xiang Peng, Feifei Li, Eric Lo [paper]
When Private Blockchain Meets Deterministic Database

Ziliang Lai, Chris Liu, Eric Lo [paper]
Top-K Deep Video Analytics: A Probabilistic Approach

Ziliang Lai, Chenxia Han, Chris Liu, Pengfei Zhang, Eric Lo, Ben Kao [paper] [code]
Everest: A Top-K Deep Video Analytics System

Ziliang Lai, Chenxia Han, Chris Liu, Pengfei Zhang, Eric Lo, Ben Kao [paper] [video]
Saguaro: Efficient Processing of Transactions in Wide Area Networks using a Hierarchical Permissioned Blockchain

Mohammad Javad Amiri, Ziliang Lai, Liana Patel, Boon Thau Loo, Eric Lo, Wenchao Zhou [paper]
Bonspiel: Low Tail Latency Transactions in Geo-Distributed Databases

Fan Cui, Eric Lo, Srijan Srivastava, Ziliang Lai [paper]