Xin Dong Zhang
Tongyi Ling Code Technical Specialist
Profile: Graduated from Columbia University, now working in AliCloud Group, engaged in long-term research and practice in the field of software engineering intelligence, including code generation, code search, defect localization, repair recommendation, test generation, etc., in ICSE, FSE Industry track has several papers published. He is now focusing on the plug-in side, service side development and model data processing and effect tuning of Tongyi Spirit Code products, and is committed to improving the quality and efficiency for the majority of developers to improve the sense of well-being of research and development.
Topic
Human-computer collaboration trends and effect optimization practices
Introduction: Big models have brought about a new evolution of human-computer collaboration, and AIGC has been integrated into all stages of software development. In this session, the impact of AIGC on software development and the current stage of human-computer collaboration model will be explained, which leads to the biggest pain point of the current collaboration model: the effect problem. Aiming at the effect problem of code completion, code Q&A and other multi-scenarios, this topic introduces cross-file perception in the reasoning phase, large-scale offline code analysis, trigger timing and generation granularity adaptation, library retrieval enhancement, enterprise knowledge base and Agent mode and other effect optimization means to improve the quality of large model generation. Outline: 1. the impact of AIGC on software development 2. The effect of AIGC on software R&D 3. Future AIGC R&D Enablement Outlook
