Changan Niu

Biography

I am currently a Yuxiu Young Scholar at the Software Institute, Nanjing University. I received my Ph.D. in Software Engineering from Nanjing University in June 2025. From July 2025 to July 2026, I worked as a Senior R&D Engineer at Huawei Consumer Business Group.

My research primarily focuses on AI for Software Engineering, including binary program understanding and decompilation, AI-driven code optimization, and code intelligence and program representation learning. I am also actively exploring embodied AI, supported by the research expertise and resources of our team. I have published several papers in top-tier international conferences and journals, and as the first author, I received the ACM SIGSOFT Distinguished Paper Award at ICSE 2024.

I welcome students interested in software engineering, artificial intelligence, and embodied AI to contact me!

Experience

• 2026.7–presentYuxiu Young Scholar, Nanjing University
• 2025.7–2026.7Senior R&D Engineer, Huawei Consumer Business Group
• 2022.10–2023.10Visiting Scholar, Singapore Management University, Singapore (Advisor: David Lo)
• 2021.9–2025.6Ph.D. in Software Engineering, Nanjing University (Advisor: Bin Luo, Chuanyi Li)

Research Interests

My research lies at the intersection of artificial intelligence and software engineering. I am interested in teaching machines to understand, generate, and optimize programs, from source code to compiled binaries, and in making such AI-driven techniques accurate, efficient, and practical. Currently, my work centers on the following directions.

• ◆ Binary Program Understanding & Decompilation
  I study the recovery of high-level program semantics from binaries without source code or debugging information, including high-fidelity decompilation, symbol and data type recovery, and program structure and behavior understanding. My research combines static program analysis, compiler techniques, and machine learning to improve the accuracy and practical utility of binary analysis and decompilation.
• ◆ AI-Driven Code Optimization
  I investigate AI-driven techniques for improving program execution efficiency while preserving functional correctness. This includes optimizing for various performance metrics such as runtime speed, memory usage, and energy consumption, and exploring methods that combine static analysis, dynamic analysis, and machine learning for intelligent code optimization.
• ◆ Code Intelligence & Program Representation Learning
  I study code representation learning, pre-training of code models, cross-task generalization, and model evaluation for software engineering tasks. My previous work includes SPT-Code and FAIR, covering program representations from source code to compiler intermediate representations.
• ◆ Embodied AI
  I am interested in embodied intelligence, particularly how intelligent agents perceive, reason, plan, and act in real-world environments. This research explores the integration of foundation models with robotic systems to bridge AI and the physical world. Our research team has committed substantial resources to embodied AI, providing students with access to industry-grade research infrastructure, practical development platforms, and internship and career opportunities through industry collaborations. These resources support both fundamental research and real-world applications.

Admissions

I welcome motivated undergraduate and graduate students interested in AI, software engineering, and embodied AI!

I am committed to providing hands-on research mentorship throughout the entire research process, from identifying research problems and reviewing the literature to designing methods, conducting experiments, writing papers, and addressing reviewer feedback. I work directly with students to help them develop strong and independent research skills. Drawing on my experience in publishing multiple first-author papers at top-tier conferences, I help students identify promising research problems, develop feasible research plans, and work toward publishing high-quality first-author papers that meet their degree requirements.

Our research team has also made substantial investments in embodied AI, offering access to industry-grade research resources, practical development platforms, and internship and career opportunities through industry collaborations. These opportunities help students pursue both academic excellence and professional development.

Whether you plan to pursue a Ph.D., build an academic career, or work on cutting-edge technologies in industry, I encourage you to reach out if you have solid programming skills, intellectual curiosity, and a strong commitment to research.

Publications

Related Links

Nanjing University | Software Institute, Nanjing University | State Key Laboratory for Novel Software Technology | Vincent Ng