RAG (Retrieval-Augmented Generation) System for Student Information
Overview
This service provides a Retrieval-Augmented Generation System (RAG) designed to support students and academic staff by integrating information from multiple institutional sources. The system enables users to query and retrieve relevant information from curated document collections containing protocol descriptions, legislation related to higher education, the UP FAMNIT faculty website, and detailed information about professors, courses, and study programs. It supports students, prospective students, and faculty staff by providing reliable, context-aware answers derived directly from official institutional documents. The service improves access to academic information, helps users navigate study regulations and administrative procedures, and provides guidance regarding courses, study requirements, and faculty services. The service uses a Retrieval-Augmented Generation approach that combines a language model with a semantic document retrieval system. Domain documents are processed and indexed to enable efficient semantic search and knowledge retrieval. The system uses Sentence Transformers to generate document embeddings with the BGE-M3 embedding model, while the embeddings are stored and searched using the Qdrant vector database. This architecture enables fast and accurate retrieval of relevant document passages that are used to generate context-aware responses.
Outcome
The system provides accurate, context-aware responses to user queries based on the integrated institutional knowledge base. Users receive explanations and references to relevant sources covering topics such as study regulations, administrative procedures, course descriptions, study requirements, and faculty services. The service improves access to academic information and supports better navigation of higher education processes.
Service Card
| Category | AI products |
| Domain | Education Technologies (EdTech) |
| Vertical | Digital Society |
| Target users | Students, Prospective Students, Academic Staff, Administrative Staff, University Student Support Services |
| Service type | Model |
| Outputs | RAG vector database with a selected off-the-shelf LLM model for training; chat system for inference. |
| Requirements | Requires NVIDIA GPU node. |
| Limitations | |
| Availability | Q4/2026 |
| Licensing | MIT and CC BY-NC-SA 4.0 |
| Visual content | |
| Visual caption |