I’m Chengran Yang (杨诚然), a Researcher at IQuest, where I work on post-training coding agents for foundation models. I received my Ph.D. from Singapore Management University (SMU) in March 2026 under the supervision of Prof. David Lo (ACM & IEEE Fellow), and continued there as a Postdoctoral Researcher before joining IQuest. My research centers on LLM Post-Training (SFT / RL) for Code Intelligence and Software Security. I work on making Code LLMs not just generate code, but generate code that is correct, efficient, secure, and well-reasoned — bridging the gap between raw generative capability and production-grade software quality. Concretely, my work spans two threads:
- (1) Enhancing Functional and Non-Functional Properties of Code LLMs — aligning models toward code correctness, runtime efficiency, security (vulnerability detection & repair), and structured reasoning via post-training;
- (2) Mining and Distilling Developer Knowledge into LLMs — leveraging SE community artifacts (Q&A forums, code reviews, documentation) to teach models human-like development reasoning and workflows.
I have published 30+ papers at top-tier SE venues (ICSE, ASE, FSE, TSE, TOSEM, etc.), including two ACM SIGSOFT Distinguished Paper Awards (ICSE 2026, as first author; ASE 2026). Prior to SMU, I received my B.Eng. from the University of Electronic Science and Technology of China (UESTC) and interned at the Institute of Computing Technology, Chinese Academy of Sciences, supervised by Prof. Jianfeng Zhan.
🔭 I’m currently exploring agentic post-training for code security and code testing — if these topics interest you too, feel free to reach out for collaboration!
News
- 🏆 Sep. 2026 “AgentExecutor: Partial Code Execution via Agentic Context Generation” received the ACM SIGSOFT Distinguished Paper Award at ASE 2026!
- 🚀 Aug. 2026 We released CyberFactory, an open-source framework for agentic cybersecurity training, together with our OpenAegis model
- 💼 Aug. 2026 Joined IQuest as a Researcher, working on coding agents for foundation models
- 📄 Apr. 2026 One paper has been accepted by ICML 2026 (as first author)
- 📄 Apr. 2026 One paper has been accepted by ISSTA 2026 (as corresponding author)
- 🏆 Apr. 2026 My First-authored ACL paper has been selected as Best Paper Candidate
- 📄 Apr. 2026 Two papers have been accepted by ACL 2026 (2*main conferences, one as first author, one as corresponding author)
- 📄 Mar. 2026 Two papers have been accepted by FSE 2026 (one as corresponding author)
- 🏆 Mar. 2026 “Think Like Human Developers: Harnessing Community Knowledge for Structured Code Reasoning” received the ACM SIGSOFT Distinguished Paper Award at ICSE 2026!
- 📄 Mar. 2026 “Out of Distribution, Out of Luck: How Well Can LLMs Trained on Vulnerability Datasets Detect Top 25 CWE Weaknesses?” accepted by ICSE 2026
- 💰 Mar. 2026 Received the Google Cloud Research Grant to support our research on code intelligence and security
- 📄 Sep. 2025 “Token Sugar: Making Source Code Sweeter for LLMs through Token-Efficient Shorthand” accepted by ASE 2025
- 📄 Jun. 2025 “Think Like Human Developers” accepted by ICSE 2026
- 📄 Jun. 2025 “APIDocBooster” accepted by ICSME 2025
- 🔥 May. 2025 “R2Vul” ranked top-3 most liked papers in SE on the Alphaxiv platform
- 📄 May. 2025 “ACECode” received a major revision from TOSEM
