Shaobo Zhang
Algorithm Researcher, CodeGeeX, Zhipu AI
He graduated with a Master’s degree in Software Engineering from the University of Texas in 2017. Since then, he has been engaged in NLP research, focusing on model training and application deployment. In 2023, he joined Zhipu AI, where he is responsible for training the CodeGeeX code model and CodeGeeX Embedding model, and has led the development and iteration of CodeGeeX Agent functionalities.
Topic
Application of LLM Agents in Software Engineering
In the field of software engineering, a core challenge has always been how to enable artificial intelligence to truly participate in real-world development and maintenance. With the advancement of large language models (LLMs), the shift toward Agents has become a key path: empowering models not only to understand code but also to autonomously plan, invoke tools, and complete complex tasks. This talk will introduce typical application scenarios of LLM Agents in software engineering, with a focus on how we leverage them to solve real-world problems in development. We will not only evaluate the capability boundaries of LLMs in software engineering but also envision the future form of AI development assistants. Outline: 1. Introduction: The integration of AI and software engineering 2. Why we need LLM Agents & how they address real development needs 3. Experiments and case studies 4. Future prospects and challenges
