Reasoning models can do impressive things. But what actually makes them different from a standard LLM?

This course gets beneath the surface.

You’ll explore the research breakthroughs that led to reasoning models and how they work under the hood. Learn to use this next level of AI models to tackle problems, compare the behavior of different systems, and examine what their outputs can — and can’t — tell you about how they arrived at an answer.

From there, you’ll dig into the techniques shaping modern reasoning AI. Prompt engineering. Context engineering. Reinforcement learning. RLHF. Process reward models. Test-time compute.

You’ll also see where things get messy.

Why can reasoning models become confused? What happens when a model produces convincing but unreliable reasoning? How can verification help separate a plausible answer from a genuinely strong one? And why don't they like cats...?

By the end, you’ll have a clearer mental model for how reasoning systems are trained, evaluated, prompted, and improved - and a more practical way to work with them.