DEVELOPMENT AND IMPLEMENTATION OF AN INTELLIGENT RAG CHATBOT FOR UNIVERSITY WEB RESOURCES WITHIN THE FRAMEWORK OF BUILDING A DIGITAL INCLUSIVE EDUCATIONAL ECOSYSTEM

Abstract

The article presents the architecture, technical implementation, and deployment results of an intelligent chatbot based on Retrieval-Augmented Generation (RAG) technology integrated into the web resources of the Simon Kuznets Kharkiv National University of Economics. The study was carried out within the framework of the comprehensive applied research project “Development of an Open Digital Inclusive Educational Ecosystem of a University to Ensure the Continuity of Higher Education in Ukraine,” funded by the general state budget. The paper provides a detailed description of the system’s multilayer architecture based on the C4 model, including a two-stage query routing mechanism (Topic-based Routing), the implementation of the ReAct Agent pattern using the LangGraph framework, a Data Ingestion Pipeline for data collection and indexing, and the integration of the Pinecone vector database with OpenAI large language models. Key architectural decisions are substantiated, including asynchronous request processing, topic-based filtering of semantic search results, and separation of responsibilities at the Django application level. The proposed approach ensures system scalability, controllability of response generation, and reduced barriers to information access for all categories of educational process participants, including individuals with special educational needs.