Juntao Chen

My research focuses on AI for Software Engineering (AI4SE), with a particular emphasis on AI-enabled recommender systems for code generation tasks.

I am an experienced software engineer with professional experience at several Fortune Global 500 companies, including Lenovo Group Limited and China Mobile Limited. Moreover, I have a strong track record of leading teams in large-scale software development projects.

I welcome connections and discussions with peers working in related research areas.

📚 Education

University of Melbourne

Apr 2026 - Apr 2030

Doctor of Philosophy (PhD) in Computer Science

Sun Yat-sen University

Sep 2021 - Jun 2025

Bachelor of Engineering in Software Engineering

🏆 Outstanding Student Scholarship, Sun Yat-sen University (2022)

📝 Publications

ICSE 2026

What to Retrieve for Effective Retrieval-Augmented Code Generation?
An Empirical Study and Beyond

Wenchao Gu, Juntao Chen, Yanlin Wang, Tianyue Jiang, Xingzhe Li, Mingwei Liu, Xilin Liu, Yuchi Ma, Zibin Zheng

In Proceedings of the 48th IEEE/ACM International Conference on Software Engineering (ICSE 2026)

Abstract PDF Github

Repository-level code generation remains challenging due to complex code dependencies and the limitations of large language models (LLMs) in processing long contexts. While retrieval-augmented generation (RAG) frameworks are widely adopted, the effectiveness of different retrieved information sources—contextual code, APIs, and similar snippets—has not been rigorously analyzed. Through an empirical study on two benchmarks, we demonstrate that in-context code and potential API information significantly enhance LLM performance, whereas retrieved similar code often introduces noise, degrading results by up to 15%. Based on the preliminary results, we propose AllianceCoder, a novel context-integrated method that employs chain-of-thought prompting to decompose user queries into implementation steps and retrieves APIs via semantic description matching. Through extensive experiments on CoderEval and RepoExec, AllianceCoder achieves state-of-the-art performance, improving Pass@1 by 10–20% over existing approaches. This study provides an experimental framework to further exploring what to retrieve in RAG-based code generation, with our replication package available at https://anonymous.4open.science/r/AllianceCoder to facilitate future research.

💼 Professional Experience

Software Engineer Intern

Jul 2024 - Oct 2024

Lenovo Group Limited, Shenzhen, China

Led the full-stack development of the NECAP Department Asset Management System. Developed the front-end with Vue.js and Element Plus, and built the back-end with Golang and the GIN framework. Implemented key features including asset inventory, asset transfer, and asset editing, ensuring smooth workflows and improving overall system efficiency for managing departmental assets.
Implemented WebSocket for real-time requests and heartbeat detection, improving stability in high-concurrency environments. Deployed the system with Nginx for reverse proxy and load balancing to optimize request handling, enhance system resilience, and improve response speed by 20% in high-concurrency scenarios.

Site Reliability Engineer Intern

Jan 2024 - Feb 2024

China Mobile Limited Fujian, Fuzhou, China

Developed an automated operations monitoring and fault diagnosis system using Python, incorporating email alerts via the smtplib package and APIs for seamless automatic work order creation.
Built a large language model (LLM) toolchain with LangChain to assist the operations team. Created various APIs with Python and the Flask framework for real-time server status updates, log retrieval, and critical data queries, which were integrated into the LangChain toolchain to improve the precision of LLM-generated outputs.

Software Engineer Intern

Jul 2023 - Aug 2023

Evecom Technology Co., Ltd., Fuzhou, China

Developed the back-end of a Legal Knowledge Question Bank System using Java, the Spring Boot framework, and MyBatis for database interactions, implementing management operations for question bank data. Integrated Redis for data caching, reducing API response time by 50%.
Implemented fine-grained access control using Spring Security and JWT for different user roles, enhancing system security and data protection. Extracted and organized legal questions from Word documents using Apache POI and regular expressions, achieving a 98% extraction success rate while effectively handling exceptions.

Database Developer Intern

Jan 2023 - Feb 2023

AsiaInfo Technologies Limited, Fuzhou, China

Designed the Oracle database for an Infrastructure Management system, developing efficient database schemas and optimizing indexing strategies to improve query performance. Additionally, implemented robust backup and recovery plans to ensure data integrity and maintain system reliability.
Configured Oracle Data Guard for disaster recovery and implemented a regular backup schedule with point-in-time recovery (PITR) to ensure data availability and integrity. Utilized Oracle Enterprise for database optimization, improving data retrieval speed of key communication traffic reports by 50%.