This chapter introduces conversational recommendation systems (CRS), focusing on the integration of reinforcement learning (RL) and large language models (LLMs) to enable dynamic, interactive recommendations. It starts by outlining foundational RL algorithms such as multi-armed bandits, deep Q-networks, and policy gradients. The chapter then discusses RL and LLM applications in dialogue management, personalization, and reward design. The chapter also details key CRS modules, including intent detection, state tracking, clarification mechanisms, and evaluation strategies. A practical tutorial demonstrates how to build a CRS using RL and LLMs, offering insights into system design, preference extraction, and reward modeling for real-world deployment.