Towards an Efficient, Customizable, and Accessible AI Tutor

Juan Segundo Hevia, Facundo Arredondo, Vishesh Kumar
Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop, PMLR 273:250-254, 2025.

Abstract

We propose a novel AI tutoring system that combines a Retrieval-Augmented Generation (RAG) pipeline with a lightweight language model to provide efficient, customizable, and accessible educational support. Designed to operate offline with minimal computational resources, the system addresses the challenges faced by resource-constrained communities. To develop its knowledge capabilities, we explore various retrieval strategies starting from a knowledge base of college textbooks. This work lays the foundation for developing adaptable and equitable AI tutoring solutions that bridge educational gaps and empower learners in under-resourced communities.

Cite this Paper

BibTeX
@InProceedings{pmlr-v273-hevia25a, title = {Towards an Efficient, Customizable, and Accessible AI Tutor}, author = {Hevia, Juan Segundo and Arredondo, Facundo and Kumar, Vishesh}, booktitle = {Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop}, pages = {250--254}, year = {2025}, editor = {Wang, Zichao and Woodhead, Simon and Ananda, Muktha and Mallick, Debshila Basu and Sharpnack, James and Burstein, Jill}, volume = {273}, series = {Proceedings of Machine Learning Research}, month = {03 Mar}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v273/main/assets/hevia25a/hevia25a.pdf}, url = {https://proceedings.mlr.press/v273/hevia25a.html}, abstract = {We propose a novel AI tutoring system that combines a Retrieval-Augmented Generation (RAG) pipeline with a lightweight language model to provide efficient, customizable, and accessible educational support. Designed to operate offline with minimal computational resources, the system addresses the challenges faced by resource-constrained communities. To develop its knowledge capabilities, we explore various retrieval strategies starting from a knowledge base of college textbooks. This work lays the foundation for developing adaptable and equitable AI tutoring solutions that bridge educational gaps and empower learners in under-resourced communities.} }
Endnote
%0 Conference Paper %T Towards an Efficient, Customizable, and Accessible AI Tutor %A Juan Segundo Hevia %A Facundo Arredondo %A Vishesh Kumar %B Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop %C Proceedings of Machine Learning Research %D 2025 %E Zichao Wang %E Simon Woodhead %E Muktha Ananda %E Debshila Basu Mallick %E James Sharpnack %E Jill Burstein %F pmlr-v273-hevia25a %I PMLR %P 250--254 %U https://proceedings.mlr.press/v273/hevia25a.html %V 273 %X We propose a novel AI tutoring system that combines a Retrieval-Augmented Generation (RAG) pipeline with a lightweight language model to provide efficient, customizable, and accessible educational support. Designed to operate offline with minimal computational resources, the system addresses the challenges faced by resource-constrained communities. To develop its knowledge capabilities, we explore various retrieval strategies starting from a knowledge base of college textbooks. This work lays the foundation for developing adaptable and equitable AI tutoring solutions that bridge educational gaps and empower learners in under-resourced communities.
APA
Hevia, J.S., Arredondo, F. & Kumar, V.. (2025). Towards an Efficient, Customizable, and Accessible AI Tutor. Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop, in Proceedings of Machine Learning Research 273:250-254 Available from https://proceedings.mlr.press/v273/hevia25a.html.

Related Material