Digital Transformation of Engineering Courses Empowered by Large Language Models

This study explores the role of educational large language models in the digital transformation of engineering courses. A four-in-one transformation framework is constructed, covering teaching content, teaching modes, practical components, and teaching evaluation. The core course, Facility Planning in the Industrial Engineering program, is selected for instructional experimentation. In terms of teaching content, a knowledge graph is generated through a large language model, and industry-oriented frontier cases are dynamically produced. In teaching modes, the instructional workflow before, during, and after class is reshaped through an AI-supported dual-teacher collaboration mechanism. In the practical component, a facility-planning system integrating the DeepSeek model and the Systematic Layout Planning (SLP) method is developed. Natural-language interaction enables the full workflow from data analysis to solution optimization, supporting intelligent design training that follows the sequence of input, analysis, generation, and optimization. In teaching evaluation, a comprehensive system is adopted by combining process-based data with multi-dimensional capability assessment. Empirical results from one semester indicate significant improvements in students’ engagement, abilities to solve complex engineering problems, and systems thinking. Meanwhile, the teacher’s role gradually shifts from knowledge transmitter to learning facilitator and context designer. The proposed framework and its implementation path may serve as a reference for digital reform in other engineering courses. Future work extends to domain-specific model development, cross-course application, and the construction of shared resource ecosystems.