Lishui Fan

flscode@zju.edu.cn

Hangzhou, China

I am a third-year Ph.D. student at the College of Computer Science and Technology, Zhejiang University, advised by Zhongxin Liu and Shanping Li. My research focuses on large language models for code generation, with broader interests in code-change understanding, coding agents, and LLM post-training. I was fortunate to intern at Taotian Future Life Lab and Douyin AI4SE, ByteDance, and I am currently interning with Baidu’s ERNIE Bot team. Over the past two years, three of my first-author CCF-A papers have been published or accepted in ACM TOSEM, ASE, and ACL, along with a first-author paper in Findings of ACL.

Code generation Code-change understanding Coding agents LLM post-training

news

May 31, 2026	Our collaborative work with Tencent, Balancing Latency and Accuracy of Code Completion via Local-Cloud Model Cascading, was accepted to FSE 2026.
May 05, 2026	ReCode: Reinforcing Code Generation with Reasoning-Process Rewards has been accepted to the ACL 2026 Main Conference.
Apr 08, 2026	ZeroCoder: Can LLMs Improve Code Generation Without Ground-Truth Supervision? is available as an arXiv preprint.
Nov 15, 2025	FGit: Fault-Guided Fine-Tuning for Code Generation was accepted to ASE 2025.
Jun 30, 2025	SEK: Self-Explained Keywords Empower Large Language Models for Code Generation was accepted to Findings of ACL 2025.

selected publications

TOSEM

Exploring the Capabilities of LLMs for Code-Change-Related Tasks

Lishui Fan, Jiakun Liu, Zhongxin Liu, and 3 more authors

ACM Transactions on Software Engineering and Methodology, 2025

CCF-A

Abs DOI

This work studies how large language models perform on code-change-related tasks, including code review generation, commit message generation, and just-in-time comment update.
ACL

ReCode: Reinforcing Code Generation with Reasoning-Process Rewards

Lishui Fan, Yu Zhang, Mouxiang Chen, and 1 more author

In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics, 2026

CCF-A Accepted to ACL 2026 Main Conference

Abs DOI

This work proposes ReCode, a reasoning-reinforced code generation framework that trains reasoning-process reward models and gates process rewards with strict execution outcomes.
ASE

FGit: Fault-Guided Fine-Tuning for Code Generation

Lishui Fan, Zhongxin Liu, Haoye Wang, and 3 more authors

In Proceedings of the 40th IEEE/ACM International Conference on Automated Software Engineering, 2025

CCF-A

Abs DOI

This work introduces fault-guided fine-tuning for code generation by emphasizing error-sensitive code segments during training.
ACL

SEK: Self-Explained Keywords Empower Large Language Models for Code Generation

Lishui Fan, Mouxiang Chen, and Zhongxin Liu

In Findings of the Association for Computational Linguistics: ACL 2025, 2025

CCF-A

Abs DOI

This work proposes self-explained keywords to help LLMs better understand low-frequency, problem-specific terms in code generation prompts.
arXiv

ZeroCoder: Can LLMs Improve Code Generation Without Ground-Truth Supervision?

Lishui Fan, Mouxiang Chen, Tingwei Zhu, and 4 more authors

arXiv preprint, 2026

Abs DOI

This work studies whether LLMs can improve code generation without ground-truth supervision through label-free co-evolution of code and test generation.
FSE

Balancing Latency and Accuracy of Code Completion via Local-Cloud Model Cascading

Hanzhen Lu, Lishui Fan, Jiachi Chen, and 3 more authors

In Proceedings of the 34th ACM International Conference on the Foundations of Software Engineering, 2026

CCF-A

Abs DOI

This collaborative work studies local-cloud model cascading for low-latency and high-accuracy code completion.