This study aims to investigate the potential of integrating advanced artificial intelligence models, specifically GPT-3.5, to enhance student support in university modules.

The project achieved the development of an enhanced chatbot that successfully leverages Retrieval- Augmented Generation to access and integrate targeted information into its responses, thereby significantly improving the specificity and accuracy of its assistance. Fine-tuning was applied to better understand and process complex academic queries specific to university courses, achieving a high degree of contextual alignment to the source lecture and response relevance.

The model was rigorously tested and validated within an educational setting, demonstrating improved performance in generating accurate responses compared to the standard GPT model. Feedback from real university students confirmed that the enhancements could substantially improve their learning experience by providing precise and informative responses.

The code repository for this project can be found here.