RAG and AI agents

Ground answers in data and let AI use tools.

📖 Key ideas

RAG (Retrieval-Augmented Generation) fetches relevant documents and adds them to the prompt.
Embeddings and vector databases power retrieval.
Tool use lets models call functions (search, calculator, database).
Agents loop: plan → act → observe until a goal is met.
Keep humans in the loop for risky actions.

💡 Real-life example: RAG is an open-book exam; agents are interns who can use office tools under supervision.

RAGRetrieval-Augmented Generation

EmbeddingVector representing meaning

Tool useModel calling external functions

AgentLLM looping with tools toward a goal