An AI-powered system for discovering novel drug-disease relationships using Relational Graph Convolutional Networks (R-GCN) and Retrieval-Augmented Generation (RAG). This project integrates heterogeneous biomedical knowledge graphs with natural language processing to predict and explain potential drug repurposing candidates.