Chi Zhang (张驰)

Master student
Computer Science, Department of Computer,
Zhongguancun Campus, Central Teaching Building 1404,
Beijing Institute of Technology,
Beijing, China
Tel: +86 18810575675
Email: zhangchi5675@163.com

[CV] [My GitHub] [Undergraduate transcript] [中文主页]

Master in Beijing Institute of Technology (2023.9 ~ 2026.6)

• Recommended for admission to Beijing Institute of Technology (double first-class A category, 985, 211)

• Department/Major: School of Computer Science, Computer Science and Technology

• Laboratory: Key Laboratory of Intelligent Information Technology, Beijing Institute of Technology

• Scholarship: Special Academic Scholarship of Beijing Institute of Technology

Undergraduate in Beijing Institute of Technology (2019.9 ~ 2023.6)

• Department/Major: School of Computer Science, Computer Science and Technology

• Rank in Major: 15/323; GPA: 3.8/4.0 [Transcript]

• Scholarships: National Scholarship (2021.10), First-class Scholarship 4 times, Second-class Scholarship 3 times, Excellent Moral Education Tutor Scholarship
  

• Honors: Outstanding Student of Beijing Institute of Technology (2020), Outstanding Student Model of Beijing Institute of Technology (2021), Outstanding Student Leader of Beijing Institute of Technology (2022), Outstanding "Three-Full Education" Moral Education Tutor (2021.11), School "Outstanding Communist Youth League Member" Honor
  

• Awards in Competitions: Bronze Award in the 12th Challenge Cup National College Students' Entrepreneurship Competition, Second Prize in Beijing Institute of Technology Mathematical Modeling Competition, First Prize in Foreign Language Teaching and Research Press National English Reading Competition, etc.

Lenovo Research NLP Algorithm Intern (2024.4 ~ 2024.7)

• Main Achievements: Assisted the team in achieving first place in the Chinese retrieval task on the MTEB leaderboard. The project involved large-scale data processing, large model fine-tuning, efficient retrieval algorithm implementation, and significant model accuracy improvement, eventually standing out among global competitors to win first place.

• My Role: Responsible for using RAG knowledge enhancement technology and LangChain to generate training data, and for data cleaning, I designed a Multi-Aspect based data generation and cleaning method. Additionally, I was in charge of optimizing the fine-tuning process of large models by incorporating Grad Cache technology into the training of large models, thereby increasing the Batch size limitations during training and enhancing the training effects of contrastive learning, ultimately improving the leaderboard results.

• Main Technologies Involved: Langchain, Large Model Fine-Tuning with Lora+Peft, Accelerate&Deepspeed, Pytorch