Smart Academic Chatbot for Student Information Services Based on Transformer-Fine Tuning

Muhammad Nazwan (Institut Teknologi Batam)
Ananda Hilmy (Institut Teknologi Batam)
Azua Ramadhan (Institut Teknologi Batam)
Fernandi Lucky Putra (Institut Teknologi Batam)
Natasya Nur Hadi (Institut Teknologi Batam)
I Gusti Bagus Agung Dharmaya (Institut Teknologi Batam)
Mhd Adi Setiawan Aritonang (Institut Teknologi Batam)
Deosa Putra Caniago (Institut Teknologi Batam)

Jurnal Desain dan Analisis Teknologi
Vol. 5 No. 2 (2026): Juli

This research developed a Transformer-based Smart Academic Chatbot using the fine-tuning method to generate more accurate responses tailored to students' needs. The development process follows the Waterfall model, starting from the collection of academic question-answer data, pre-processing, architecture design, to implementation and testing. The evaluation results show that the chatbot is able to provide relevant answers with a good level of accuracy, as well as improve the ease of access to academic information. This system has the potential to be a smart solution in supporting the digital transformation of campus academic services.