Design and Development of a Generative AI-Assisted Learning Software System Architecture Tool

Abstract

As software systems become increasingly complex, software engineering employs standardized development processes to enhance efficiency, ensure software quality, and reduce maintenance costs. Among these, the completeness and accuracy of requirement specifications significantly impact the final product. In recent years, the application of generative artificial intelligence (AI) in software engineering has garnered growing attention, offering assistance in requirement analysis and accelerating code generation to improve development efficiency. However, current generative AI models face challenges such as inconsistent code quality and difficulties in handling large-scale software architectures. This study proposes an approach that integrates function-level code generation, software frameworks, and natural language requirement specifications. By optimizing function-level generation and architectural structuring, the proposed method enhances the accuracy and stability of AI-generated code, shortens development cycles, and strengthens developers’ capabilities in planning large-scale software systems, ultimately improving the overall efficiency of software engineering.