Features of Using Frameworks and Artificial Intelligence Language Models for JavaScript Code Optimization in Web Application Development

This work is dedicated to exploring methods for writing rational and functional code when developing web applications using JavaScript frameworks, as well as paths for its optimization. The results of JavaScript code optimization without losing upstream code functionality using artificial intelligence (AI) models based on Large Language Models (LLMs) are presented. The essence of the experiment involved loading a JavaScript file from a web application package for recognition and incremental optimization in a dialogue mode conducted by various AI models trained on diverse programming language datasets. During recognition, the AI identified the main scripts by their purpose and optimized the code as required. Using GPT-4o, an improved version of the code was obtained, featuring changes such as removal of redundant checks and operations, consolidation of similar operations into functions, elimination of code duplication, and others. As a result, code readability and performance, as well as page loading speed, were improved. Similarly, other AI models were tested, comparing their capabilities in both optimization and generation of code with specified functionalities.