Hossain, Muhammad IqbalShaan, Md. MahinuzzamanTahsin, NuzhatPolock, Afridi AlamEmu, Lamia AlamMehreen, Nafisa2025-09-032025-09-0320252025-06ID 21101208ID 21301206ID 21301303ID 21301662ID 23241114http://hdl.handle.net/10361/26645Cataloged from PDF version of thesis.Includes bibliographical references (pages 58-61).This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2025.The core contribution of this paper lies in con structing and employing a high quality, Bangla-language dataset specifically for educational settings. While AI and NLP have transformed learning environments elsewhere in the world, Bangla, as the seventh most spoken language, remains under represented, particularly in educa tion. This piece addresses the lack by focusing on constructing curriculum-aligned dataset that comes in the form of Bangla context paragraphs and corre sponding question-answer pairs. These sources of data are mined and cleansed from national curriculum textbooks and teach ing resources, ensuring linguistic appropriateness and academic usefulness. The cleaned dataset is poised to facilitate natural lan guage processing algorithms bespoke to Bangla, enabling the development of smart systems that can successfully understand, process, and generate context-dependent responses appropriate to education settings. With this work, we intend to clear the way for more available, localized AI-powered learning environments that are accessible for Bangla-speaking students, especially in low-resource environments.61 pagesenBRAC University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.DatasetNatural language processingLarge language modelTokenizerBangla-speaking studentsArtificial intelligenceNatural language processing (Computer science).Artificial intelligence.Real time--Data proccessing.Thesis