International Journal of Scientific Research and Engineering Development

Colleges and other higher education institutions are increasingly dependent on online resources for academic and administrative data. Although these technologies combine institutional resources, they frequently employ keyword-based search methods and predetermined content formats, which restricts the user's ability to access information rapidly. As a result, students occasionally put off looking for the information they need. These architectural constraints also limit typical rule-based chatbots that are used on these platforms, and that rely on predetermined patterns and a lack of in-depth semantic understanding. As a result, their academic achievement is negatively impacted by their inability to comprehend questions that are context-dependent or use dynamic language. In order to address these issues, this study suggests integrating a Smart Campus Conversational Assistant onto a university website to enable intelligent, domain-specific access. Semantic vector-based retrieval and regulated generative language modelling are combined in the system's Retrieval-Augmented Generation (RAG) design. Before being indexed by Facebook FAISS-powered semantic vector search for speedy vector similarity matching, the Sentence Transformer model preprocesses and encodes institutional knowledge into dense vector representations. In order to find relevant information, incoming queries are buried inside the same vector space throughout runtime. The results are then sent to a Large Language Model (LLM) for logical and well-documented responses. The suggested hybrid retrieval-generation method increases the level of semantic relevance as well as context alignment through a reuction in false positives, thereby enhancing the transmission of reliable information through intelligent campus networks.

Zoya Anjum, Smanavi Reddy, Geetha Sri, Bazila Wahaj,"Smart Campus Conversational Assistant An Intelligent System for College Queries" International Journal of Scientific Research and Engineering Development, V9(2): Page(1805-1811) Mar-Apr 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.