Development of a Chat Assistant Using Large Language Models for Personalized Mathematics Tutoring to Overcome Educational Disparities in Sri Lanka

The accessibility and quality of education in Sri Lanka face significant disparities, particularly between rural and urban areas. This research developed a personalized Intelligent Tutoring Assistant (ITA) powered by large language models (LLMs) to address this challenge. Focusing on the Grade 9 algebraic expressions lesson, the ITA was designed to provide personalized mathematics tutoring to Sri Lankan students. By aligning with the Sri Lankan school syllabus and incorporating iterative teacher feedback, the ITA was refined to improve its educational effectiveness. The ITA demonstrated its potential as a valuable supplementary tool that bridges classroom instruction and personalized student support. This scalable solution can address diverse learning needs and fill gaps in traditional education. Its applications extend beyond mathematics education, with potential across STEM disciplines and various educational contexts. The ITA has the potential to evolve into a comprehensive educational tool, promoting equitable access to quality education, especially in resource-constrained settings.