about_img

About me

Master Student @ HITSZ
Research Areas: Large Language Models, Security, Federated Learning

Profile
  • School：

    Harbin Institute of Technology, Shenzhen

  • Birthday:

    May 2000

  • Email: muxichenz@outlook.com
  • Address:

    HITSZ, University Town, Shenzhen

Publications

  • image
    CCS'24 Published

    Unveiling Vulnerabilities of LLMs' Split Fine-tuning

  • image
    Arxiv

    Centaur: Privacy-preserving Transformer Inference Framework

  • image
    SSRN Preprint

    Large Language Models Federated Learning Toolkit

  • image
    TIST'23 Published

    A Survey on Trustworthy Federated Learning

  • image
    Arxiv

    FedCVD: The First Real-World Federated Learning Benchmark on Cardiovascular Disease Data

Work Experience

Including internships from both undergraduate and master’s periods.

Pengcheng Lab

2019.12-2020.5 AI Department, Intern

Open-source community intern. Responsible for researching top papers and maintaining the open-source code hosting platform iHub.

ByteDance

2021.1-2021.8App Framework, RD Intern

In my junior year, I worked as a product RD in the CloudService team of ByteDance, responsible for the functional iteration of the product DataRangers-AndroidSDK

SenseTime Tech.

2022.6-2022.8 Backend RD Intern

Participated in building a data platform, including debugging and deploying visual models, as well as developing related services.

Shanghai Academy of AI for Science

2024.10- Intern

Participated in a large model privacy protection project, responsible for designing data extraction attack strategies for large language models and developing attack toolkits.

Interests & Skills

Kinda Broad

LLM Security

Research on privacy and security of large language models, covering privacy leakage and protection in distributed learning frameworks, data retention and extraction in large models, and jailbreak studies of large models.

Federated Learning

Research on federated learning, including federated graph learning, applications of federated learning in the medical field, and secure multi-party federated learning frameworks.

Mobile Development-kotlin/java

Native Android app developer with experience in Java/Kotlin, a preference for MVVM architecture, and Jetpack Components. Previously interned as an Android developer at BD.

Backend developer experienced with Go, JavaScript, and Java.

Experienced in backend development with Golang, Node.js, and Java, with stronger familiarity in Node.js and Go, and experience using MySQL and Redis.

Game Development-C++

In sophomore year, I developed a trading card game similar to Slay the Spire using C++ and the Cocos2d-x engine.

UI/Video Design

Enjoy working on UI and video design, including demo videos for projects like the HITSZ Assistant and CloudLight.

Engineering Projects

Finished during my undergraduate study

17

Awards Since Undergraduate Studies.

7000

Total Project Users.

2

Software Copyright.

+++

Github stars

Awards & Honours

Fangbinxing Scholarship

Ranked first

A total of 4 postgraduate students rewarded

Master’s Academic Scholarship.

2023-2024 First-class (Top 10%)

2024-2025 First-class (Top 10%)

Outstanding Graduate

2023 HITSZ

Outstanding Graduate of the 2023 Class, Harbin Institute of Technology, Shenzhen.

Outstanding Volunteer

2019-2020 Awarded by the School of Economics and Management, Harbin Institute of Technology (Shenzhen)

First-Class Project Award.

2019 HITSZ Freshman Year Annual Project Award.

The HITSZ Assistant project, which I was responsible for, excelled in completion, technical complexity, and user scale, earning the First-Class Annual Project Award.

National First-Class Award.

2020 CAAI Innovation and Entrepreneurship Competition National First-Class Award.

Leveraging the HITSZ Assistant campus community platform and its vast user base, combined with AI technology, we implemented features such as sensitive information filtering and public opinion monitoring.

Team Excellence Award.

2021 National College Student Software Innovation Competition

“Edge-Cloud Integrated Artificial Intelligence Innovation,” the CloudRise Accessible Social Platform, primarily provides disability assistance services based on computer vision technology.

Undergraduate Academic Scholarship

2020-2021 First-Class (Top 5%)

2019-2020 Second-Class (Top 10%)

2018-2019 Second-Class (Top 10%)

Outstanding Student

2020-2021 First-Class

2019-2020 Second-Class

2018-2019 Second-Class