Students Build RAG-Powered Q&A Platform to Preserve Campus Knowledge at LASU

A team of students at Lagos State University (LASU) has developed 'AMA' (Ask Me Anything), a platform designed to preserve and share student experiences that would otherwise be lost across WhatsApp groups and social media. The project was submitted for the Gemma 4 Challenge and uses Google's Gemma 4 model via OpenRouter as its answer-generation engine. The system is built on Retrieval-Augmented Generation (RAG), allowing answers to update in real time as new student stories are submitted without requiring model retraining. Student submissions are converted into embeddings using Google's Gemini embedding model and stored in MongoDB Atlas, which handles vector similarity searches to retrieve relevant context for each query. The frontend is built with Next.js, Tailwind CSS v4, Zustand, and TanStack React Query.