Student AI Personal Assistant

Abstract

This work presents the Student AI Personal Assistant Using Multimodal Retrieval-Augmented Generation (RAG), a student-oriented application designed to simplify access to academic information stored in PDF documents and lecture materials. The documented system includes a React and Vite frontend and a FastAPI backend, with user authentication, PDF upload and processing, document and unit indexing, figure metadata, and document-specific conversational question answering. The intended workflow retrieves relevant academic content and uses it as context for generating responses. The project aims to make academic information easier to access while supporting source traceability. Further evaluation is required to measure retrieval relevance, answer groundedness, response latency, multimodal capabilities, and security. This manuscript describes the documented implementation and does not claim independently verified performance results.