Yushuai Ji

Postdoctoral Researcher, Department of Computer Science, Hong Kong Baptist University

香港浸会大学计算机科学系博士后研究员

Biography

I am currently a postdoctoral researcher in the Department of Computer Science at Hong Kong Baptist University, working with Prof. Jianliang Xu. I received my Ph.D. in Electronic Information from the School of Computer Science at Wuhan University, where I was advised by Prof. Sheng Wang and Prof. Zhiyong Peng in the Totem Database Lab.

我目前在香港浸会大学计算机科学系从事博士后研究，师从徐建良教授。我在武汉大学计算机学院获得电子信息博士学位，博士期间师从王胜教授和彭智勇教授，并在图腾数据库实验室开展研究。

My research focuses on large language model inference infrastructure and databases for AI. I am particularly interested in LLM inference, multi-agent data systems, multi-agent systems, and vector databases, with an emphasis on efficient inference and retrieval, hybrid vector search, approximate nearest neighbor search, and scalable data management. My work aims to build efficient data management and retrieval infrastructure for LLM and multi-agent applications.

我的研究主要聚焦于大模型推理基础设施和面向人工智能的数据库（databases for AI）。我主要研究大语言模型推理、多智能体数据系统、多智能体系统和向量数据库，并重点关注高效推理与检索、混合向量检索、近似近邻搜索以及可扩展的数据管理。我的工作旨在为大语言模型和多智能体应用构建高效的数据管理与检索基础设施。

Research Interests

• Multi-agent systems and agentic data management.
• Large language model infrastructure, inference systems, and efficient serving.
• Hybrid vector search, vector databases, and approximate nearest neighbor search.
• Scalable indexing, clustering, and machine learning for database systems.

• 多智能体系统与智能体数据管理。
• 大语言模型基础设施、推理系统与高效服务。
• 混合向量检索、向量数据库与近似近邻搜索。
• 面向数据库系统的可扩展索引、聚类与机器学习。

News

• 2026Our paper Highly-Efficient Large-Scale k-means with Individual Fairness was published in PVLDB 19(5). We develop TKM and FastTKM to improve fairness and scalability in large-scale clustering.
• 2026Our preprint Graph-centric Cross-model Data Integration and Analytics in a Unified Multi-model Database introduces GredoDB, a unified system for graph, relational, and document data.
• 2026Our work UnIS was accepted to ICDE 2026, addressing fast, updatable, and auto-selected on-device search.
• 2025Our paper Federated and Balanced Clustering for High-dimensional Data was published in PVLDB 18(11), studying balanced clustering in federated settings.
• 2025Our paper On Simplifying Large-Scale Spatial Vectors appeared at ICDE 2025, presenting a fast, memory-efficient, and cost-predictable approach to k-means.

• 2026我们的论文 Highly-Efficient Large-Scale k-means with Individual Fairness 发表在 PVLDB 19(5)。该工作提出 TKM 和 FastTKM，以提升大规模聚类中的公平性与可扩展性。
• 2026我们的预印本 Graph-centric Cross-model Data Integration and Analytics in a Unified Multi-model Database 提出了 GredoDB，一个统一支持图、关系和文档数据的系统。
• 2026我们的工作 UnIS 被 ICDE 2026 接收，面向快速、可更新和自动选择的端侧搜索。
• 2025我们的论文 Federated and Balanced Clustering for High-dimensional Data 发表在 PVLDB 18(11)，研究联邦场景下的平衡聚类问题。
• 2025我们的论文 On Simplifying Large-Scale Spatial Vectors 发表在 ICDE 2025，提出了一种快速、节省内存且成本可预测的 k-means 方法。

Selected Publications

2026

1. Highly-Efficient Large-Scale k-means with Individual Fairness.
   Shengkun Zhu, Jinshan Zeng, Yuan Sun, Sheng Wang, Yiming Wang, Yushuai Ji, Feiping Nie, Xiaodong Li, and Zhiyong Peng.
   PVLDB 2026, 19(5): 808-821. CCF-A
2. Graph-centric Cross-model Data Integration and Analytics in a Unified Multi-model Database.
   Zepeng Liu, Sheng Wang, Shixun Huang, Hailang Qiu, Yuwei Peng, Jiale Feng, Shunan Liao, Yushuai Ji, and Zhiyong Peng.
   arXiv:2603.01598, 2026. Preprint
3. Updatable Balanced Index for Fast On-device Search with Auto-selection Model.
   Yushuai Ji, Sheng Wang, Zhiyu Chen, Yuan Sun, and Zhiyong Peng.
   ICDE 2026, to appear. CCF-A

2025

1. Federated and Balanced Clustering for High-dimensional Data.
   Yushuai Ji, Shengkun Zhu, Shixun Huang, Zepeng Liu, Sheng Wang, and Zhiyong Peng.
   VLDB 2025, 18(11): 4032-4044. CCF-A
2. On Simplifying Large-Scale Spatial Vectors: Fast, Memory-Efficient, and Cost-Predictable k-means.
   Yushuai Ji, Zepeng Liu, Sheng Wang, Yuan Sun, and Zhiyong Peng.
   ICDE 2025, 863-876. CCF-A

Education

• Ph.D. in Electronic Information, School of Computer Science, Wuhan University, China. Advisors: Prof. Sheng Wang and Prof. Zhiyong Peng, Totem Database Lab.
• M.A. in Statistics, Washington University in St. Louis, USA.
• Zhangzhidong Class, Huazhong Agricultural University and Wuhan University, China.

• 电子信息博士，武汉大学计算机学院，中国。导师：王胜教授、彭智勇教授，图腾数据库实验室。
• 统计学硕士，圣路易斯华盛顿大学，美国。
• 张之洞班，华中农业大学与武汉大学，中国。

Work Experience

• CurrentPostdoctoral Researcher, Department of Computer Science, Hong Kong Baptist University, Hong Kong, China. Advisor: Prof. Jianliang Xu.
• Apr. - Jul. 2021Software Engineer, Alipay, China.
• Jan. - Apr. 2021Intern, Best Buy, USA.

• 目前博士后研究员，香港浸会大学计算机科学系，中国香港。导师：徐建良教授。
• 2021年4月 - 7月软件工程师，支付宝，中国。
• 2021年1月 - 4月实习生，Best Buy，美国。