Whenever you're ready, walk me through how you'd approach this.
How does RAG (Retrieval Augmented Generation) work, and how would different chunk sizes affect response generation?
