About Me

Intelligent data makes computing great.

I am a third-year undergraduate student in the Hongyi Honor Class, School of Computer Science at Wuhan University. I joined the Totem Database Laboratory in my freshman year, where I am fortunate to be advised by Prof. Sheng Wang. And I really appreciate the guidance from Dr. Zhiyu Chen at Amazon.

My research focuses on Systems for ML, aiming to make efficient and "intelligent" data management in modern systems. I'm passionate about and potential to research topics on AI infrastructure and Database Systems.

I also share technical notes and reflections on research through my Zhihu blog and Xiaohongshu. Feel free to visit if you use Chinese social media.

Education

Wuhan University Sept. 2024 - Present
B.Eng. in Computer Science (Hongyi Honor Class)

Experience

TOTEM Database Laboratory of WHU Jan. 2025 - Present
Research Assistant
LINs Lab, Westlake University May. 2026 - Sep. 2026
Visiting Student

Research Interests

LLM Serving System KV-cache Management Agent Memory Database System Robotics Systems

Awards and Honors

Outstanding Student, Wuhan University Apr. 2026
Active Participant in Social Activities, Wuhan University Mar. 2026
Scholarship for Excellent Academic Performance (Tianyuan Dic) Nov. 2025
Scholarship for Excellent Academic Performance, WHU Sept. 2025
Outstanding Student, Wuhan University Sept. 2025
Tencent Open Source Core Contributor / Award Sept. 2025
Outstanding Freshman Award, Wuhan University Sept. 2024

Service

Captain, WHU School of Computer Science Table Tennis Team Sept. 2025 - Jun. 2026
Led the school team in table tennis training and competitions.
Team Gold Medal Silver Medal
Individual Gold Medal Silver Medal Bronze Medal
Lecture Notes 武汉大学 · 计算机学院
记录武汉大学计算机学院课程学习中的知识整理、课程思考与实践总结,所有内容均使用中文,并按照课程划分为独立系列。
计算机组成与体系结构 计算理论导引 机器学习 C++ 程序设计 39 篇笔记
PhD Application Guide Sep. 2026
面向本科新生的博士申请与科研成长规划,讨论研究潜力、精力分配、科研积累、推荐信与长期发展。
PhD 申请 本科科研 经验分享

Research

My recent research topics focus on efficient context management in agent serving systems, efficient agent-database interaction, and agent memory

Research & Project Map

Systems, Architecture & Infrastructure for ML/AI

Hover to explore a relationship. Click a project to jump to its details.

Shared systems principles: scheduling, memory, and data movement

Select a research or project block to inspect its relationship.

CacheBridge: Efficient Cross-Model KV Cache Transfer arXiv 2026
Cross-Model KV Cache Transfer:
  • Attention-aligned, head-wise mapping across model architectures.
  • Memory-efficient GPU construction and low-cost online application.
PRISM: Fast Online Serving via Scheduling-Memory Co-design arXiv 2026.05
Online LLM Serving Optimization:
  • KV-Cache Management Optimization.
  • Scheduling-Memory Co-design for lower latency.
OpenTenBase Migration & Optimization Core Contributor 2025
Selected for the Tencent Rhino Bird Open Source Program.
  • Developed migration scripts to update OpenTenBase from v2.6 to v5.0 with zero data loss.
  • Successfully achieved the first full compilation of OpenTenBase v5.0 on Ubuntu.
  • Conferred as "Tencent Open-source Core Contributor".
OpenTenBase Migration and Compilation
XRDB: A SQL Database Engine in Rust. Starstruck 2026
A project to learn the language Rust. Key components include:
XRDB Overview

Publications

CacheBridge: Efficient Cross-Model KV Cache Transfer [Paper]
Xingyu Qu, Siyuan Lu, Zhiyu Chen, Sheng Wang, and Tao Lin
arXiv preprint (arXiv'26)

PRISM: Fast Online LLM Serving via Scheduling-Memory Co-design [Paper]
Xingyu Qu, Tianhao Lin, Yiqi Li, Zhiyu Chen, and Sheng Wang
arXiv preprint (arXiv'26)

Recruiting

Our lab at WHU always looks for high-motivated undergraduate students and interns/RA. If you are interested in topics about databases/LLM/AI infra/ML, feel free to contact Prof. Sheng Wang or email me directly. We can offer guidance and support within our means.

My experience in the lab has been highly rewarding. I have received excellent guidance and benefited from a supportive and collaborative lab environment. Along the way, I have developed practical insights into getting started with research in machine learning and database systems, and I am always open to research collaborations.