Events

Coding agents changed how a lot of people work, and most students have only seen them through a chat box. In this workshop we use one the way engineers do: in the terminal, on a real repository, with tests.

We start from an empty folder and build a small but complete project in an hour. You give the agent a plan, you delegate, you verify. We stop each time it does something wrong and we look at why.

You leave with a working project, a mental model of what an agent can and cannot do, and a checklist for using one on your own projects.

What is actually inside a language model? Tokens, embeddings, attention, a training loop. In one hour we walk through each piece with small, readable examples, so that the whole picture holds together.

We look at what changes when you scale up, data, parameters, compute, and why the large models behave the way they do. No framework magic, no black box.

From scratch format: more explanation than the Tools workshops, with the code on screen the whole time and the material published afterwards.