llm, rag, reinforcement based value acquisition, fine tuning