AI-Driven Software Design and Applications: Construction of Personalized Learning Measurement Models Under the Taxonomic Development Perspective

To address the demand for differentiated learning resources, this study designed and developed an AI-driven software using the PyCharm platform. The research framework consisted of six key stages. First, a composite function was designed to evaluate students’ academic foundations and learning capabilities, integrating historical course performance, classroom engagement, and teacher assessments. Second, the Qwen AI model was pre-trained on the Alibaba Cloud platform using four sets of domain-specific datasets. Third, a PC-based educational software was developed using PyCharm, with API integration enabling seamless communication between the local system and the Qwen AI model. Fourth, Student Baseline Data (SBD) were input into the software as feedforward information. Fifth, the API communication protocol was optimized to guide the AI model in incorporating feedforward data, thereby generating personalized responses for different user profiles. Finally, the software was deployed in a real classroom setting to collect user feedback. A controlled experiment was conducted with a control group (n=20) and an experimental group (n=17). Evaluation results showed that the experimental group achieved a mean score improvement of 0.0452 (experimental: 0.8847 vs control: 0.8395), with variance ratios further optimized from 0.0025 to 0.0039. Statistical significance was confirmed by t-test analysis (t=-2.4622, p=0.0189). The software demonstrated fast response times, accurate content delivery, and highly personalized tutoring capabilities. This study highlights the transformative potential of AI in education, particularly in delivering customized learning experiences, and contributes a practical framework for integrating AI into vocational pedagogy.