一句话：提出博士研究计划，研究让生成式 AI 智能体输出适配公司标准与开发者偏好的方法与工具，促进其在软件工程工作流中的集成。
将生成式 AI 工具集成到现有软件工程工作流时，基础模型生成的输出往往未适配特定公司标准与开发者个人偏好，导致许多公司难以在各类工作流中采用 AI 辅助软件工程。该博士项目计划在未来三年研究多种方法与工具，使 GenAI 输出能够同时考虑公司标准与开发者偏好，实现定制化与个性化。其目标在于降低 GenAI 与现有软件工程工作流之间的集成障碍，并提升其在不同软件工程活动中的采用率。摘要未给出阶段性结果或具体评估，属于研究计划与愿景阐述。
One of the hurdles of integrating generative AI (GenAI) tools into existing software engineering workflows is that the outputs generated by the foundational models are not tailored to specific company standards and developers’ individual preferences. This challenge has prevented many companies from adopting AI-assisted software engineering across their workflows. This PhD project aims therefore at investigating in the next three years different methods and tools for providing tailored and personalized GenAI outputs taking both company standards and developers’ preferences into account. The objective is to facilitate GenAI integration into existing software engineering workflows and boost adoption across different software engineering activities.