Архітектура серверної системи ші-помічника на основі RAG-підходу та векторного пошуку

Abstract

The paper presents the architecture of a specialized AI-assistant information system that implements the RetrievalAugmented Generation (RAG) approach for working with large collections of textual documents. The specialized AI assistant employs vector search to provide accurate and context-aware responses to user queries within a defined subject domain. The server-side component (backend service) is implemented using modern technologies, including Python,