Domanda di colloquio di ADDO AI

How does a RAG pipeline work? On what basis do we perform chunking?