Intelligent Teaching Assistant using RAG and AIML
International Journal of Advanced Research in Science Communication and Technology · 22.11.2026 · 368
DOI: 10.48175/ijarsct-40944 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 22.11.2026. Verfasst von Prof. S. Y. Mandlik und 3 weiteren Autor:innen. Die Arbeit nennt 15 Literaturangaben.
Abstract (Originalsprache)
Engineering students accumulate large collections of lecture notes, presentations, lab manuals and previous-year question papers, yet locating an accurate answer inside them remains a slow, manual task. General-purpose Large Language Models (LLMs) answer fluently but depend on static training data, may hallucinate facts, and do not use institution-specific material. Retrieval-Augmented Gener-ation (RAG) addresses these limitations by grounding generation in an external, updatable knowledge base. This review examines three recent IEEE studies, namely a systematic literature review of AI-driven Intelligent Tutoring Systems (ITS) in engineering education, a survey of RAG for LLMs, and a practi-cal AI-RAG architecture for real-time retrieval, and combines their findings with the project synopsis and presentation to propose an integrated framework for an Intelligent Teaching Assistant. The frame-work connects role-based document ingestion, chunking and embedding, query rewriting, hybrid (BM25 plus vector) retrieval, semantic re-ranking, LLM answer generation, Self-RAG verification and citation-based answers with AI quiz generation, notes generation and learning analytics for both students and teachers. The review identifies major research gaps: ITS studies rarely use RAG, RAG studies rarely target education, question-answering tools are often not personalized, and instructor-facing analytics are scarce. It also discusses challenges involving retriever-generator alignment, knowledge conflicts, la-tency, knowledge-base quality, privacy and evaluation, and outlines an evaluation plan based on the RAG evaluation triad. The resulting workflow, Ingest, Chunk, Embed, Retrieve, Re-rank, Generate, Verify, Cite and Analyze, provides a unified perspective for building grounded, transparent and personalized AI teaching assistance..
Zitieren
APA
Mandlik, P. S. Y., Kharmale, A. R., Mule, P. P., & Dattatray, A. G. (2026). Intelligent Teaching Assistant using RAG and AIML. International Journal of Advanced Research in Science Communication and Technology, 368. https://doi.org/10.48175/ijarsct-40944
Vancouver
Mandlik PSY, Kharmale AR, Mule PP, Dattatray AG. Intelligent Teaching Assistant using RAG and AIML. International Journal of Advanced Research in Science Communication and Technology. 2026:368. doi:10.48175/ijarsct-40944
BibTeX
@article{Mandlik2026,
author = {Mandlik, Prof. S. Y. and Kharmale, Atharva Ramesh and Mule, Prasad Pramod and Dattatray, Argade Ganesh},
title = {Intelligent Teaching Assistant using RAG and AIML},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {368},
doi = {10.48175/ijarsct-40944}
}
EUVIMED