Maximal Format-Free Record Repair was accepted to the IEEE/ACM International Conference on Automated Software Engineering (ASE) 2026.
Researcher in Code Intelligence
Zijian Luo aka Jack
MPhil graduate from The University of Sydney, soon joining Monash University.
I am currently a Research Intern at Microsoft DKI Research Group, working on code migration under the guidance of Dr. Pengfei Gao. Previously, I completed my MPhil in Computer Science at The University of Sydney under the supervision of Dr. Rahul Gopinath, and worked on supply-chain reliability at Oracle Labs under the supervision of Dr. Behnaz Hassanshahi. I am preparing to continue my research at Monash University under the supervision of Dr. Yongqiang Tian. My research interests include Code Intelligence, Agents, and Code Verification.
News
Latest Blog
More Blog PostsSoftware Engineering Research Is Dead
Why LLMs gave software engineering research a second life, and why the AI frontier may be moving beyond code.
Read postEducation
Doctor of Philosophy (PhD) in Computer Science
Oct 2026 – PresentMonash University (Melbourne, Australia)
Master of Philosophy (MPhil) in Computer Science
Mar 2024 – Mar 2026The University of Sydney (Sydney, Australia)
Bachelor of Engineering (B.Eng.) in Software Engineering
Sep 2018 – Jun 2022Hebei University (Hebei, China)
Work Experience
Research Intern
Mar 2026 – PresentMicrosoft Research Asia (MSRA), Microsoft
- Enhanced Copilot's cross-platform code migration capability by integrating a knowledge base to model platform-specific APIs, dependencies, and migration mappings.
- Built a dependency-upgrade benchmark for long-horizon software engineering tasks to evaluate MAI Coder's end-to-end planning and execution across multi-step workflows; the corresponding paper is under review.
- Collaborated with GitHub to propose leveraging coding-agent development trajectories as process-aware context for Copilot Code Review (CCR) to better understand code changes; the work is under review.
Research Assistant
May 2025 – Oct 2025Oracle Labs, Oracle Corporation (Brisbane, Australia)
- Conducted research on software supply-chain reliability by integrating LLMs with the Macaron framework, reproducing build artifacts for over 120 open-source repositories from Libraries.io with an 80% success rate.
- Designed LLM-assisted workflows for defect reproduction and static-analysis validation.
Publications
Update from Hell: Can Coding Agents Survive Hidden Breakage in Dependency Upgrades?
arXiv preprint, 2026.
Automatic Data Repair without Format Specifications
IEEE International Symposium on Software Reliability Engineering (ISSRE) 2025.
Assessing Reliability of Statistical Maximum Coverage Estimators in Fuzzing
IEEE International Conference on Software Maintenance and Evolution (ICSME) 2025.
Maximal Format-Free Record Repair
IEEE/ACM International Conference on Automated Software Engineering (ASE) 2026.