Jinhua Du¶
English
中文
Researcher · Large Language Models · Knowledge Engineering
Ph.D. in Computer Science and Technology, Tsinghua University | KEG Lab, Class of 2022
Advisor: Jie Tang
Research focuses on Machine Learning, Big Data, and Natural Language Processing (Large Language Models)
Curriculum Vitae
Read or download my CV
Choose the English or Chinese edition for a concise overview of my education, research, projects, and honors.
Current focus
Postgraduate research
LLM reasoning, medical intelligence, publications and honors
Foundations
Undergraduate experience
Machine learning, robotics, competitions and practice
Timeline
Research history
A year-by-year and month-by-month record of research progress
Research Interests¶
- Meta-Learning and Transfer Learning for Large Models: Investigating methods for transferring and adapting learning strategies across different tasks and domains
- Adaptive and Interpretable AI Systems: Exploring how to build intelligent systems that flexibly adapt to environmental changes and are more transparent and interpretable to human users
Honors & Life¶
- National Scholarship recipient, Outstanding Undergraduate Graduates of Beijing
- Fitness enthusiast and fitness content creator; served for two years as student assistant to the Secretary-General of the Beijing University Students Bodybuilding Association; 14 years of running experience; enjoys a wide range of sports; placed fifth in the Tsinghua University Bodybuilding Competition
Representative Work¶
- GLM-5 advances foundation models from vibe coding to agentic engineering; GLM-4.5 unifies agentic, reasoning, and coding capabilities; MalayGLM is a sovereign language model developed for Malaysia's local ecosystem.
- AML is the Advanced Machine Learning course website; AML-LLM is a textbook project covering the foundations, advances, and applications of LLMs; LLMDailyDigest curates daily papers, open-source frameworks, datasets, and industry updates.
- Med-Eval is a medical LLM evaluation benchmark; AiMed supports Chinese medical knowledge Q&A, paper reading, and literature retrieval; MedGPT explores LLM applications in healthcare.
- THU53 is the official website of Tsinghua University's CS Class 53 graduate student Party branch.
Basic Information¶
| Item | Content |
|---|---|
| GPA | 3.95 / 4.00 |
| Overall Rank | 1 / 329 |
| Date of Birth | 2000.04.01 |
Links¶
| Item | Content |
|---|---|
| dujh22@mails.tsinghua.edu.cn | |
| GitHub | github.com/dujh22 |
| Scholar | Google Scholar, Baidu Scholar, Aminer |
| Databases | ACL, DBLP, ORCID, IEEE, Semantic Scholar |
| Blog | CSDN |
Skills¶
| Category | Details |
|---|---|
| LLM | Proficient in the full pipeline of large model technologies including pre-training and post-training (fine-tuning, alignment) |
| AI | Research experience in AI numerical computing, robotics and autonomous driving, big data, natural language processing and large models; participated in product development for healthcare, public security, art, gaming, and sports domains |
| Programming | Proficient in Python, C/C++; familiar with Java, Lua, Go and other languages; experienced in numerical analysis and computation with MATLAB |
| English | CET-4 (580), CET-6 (525), IELTS (7.0) |
| Teaching | Certified Teaching Assistant at Tsinghua University; served as teaching assistant for graduate courses |