Module 11 · Understanding LLMs from scratch

Capstone project

Use everything from the course to build an AI project on a topic of your own choosing, start to finish on your own: set what done looks like, build something usable, show how good it is with evaluation, and finally write it up clearly for others.

Lessons

01
Choose a topic, finish it, write it up
The complete capstone process: how to choose a topic you can finish, how to set what done looks like and build the evaluation set first, how to get to a first usable version in small steps, and how to write a README people want to read and can reproduce.
2–4 weeks · Advanced

The last module has only one lesson, but it takes the most time: choose a topic yourself and complete a project from start to finish.

In the previous ten modules, every lesson had code written and results already run, and by following along you got the same output. Real projects aren't like that: nobody tells you whether to use RAG or an agent, you find the data yourself, you write the evaluation set yourself, and when results are poor you have to work out what to do. This module has you go through all of that in a project small enough to finish, but entirely real.

Lessons

Choose a topic, finish it, write it up

You're done when

You have an AI project whose topic you chose and which you finished yourself, and others can run it on their own computers by following the README.
The project has an evaluation set that runs with one command, and the README states real evaluation results, including what it does badly.
In a few minutes you can explain clearly to someone else what problem it solves, why it's designed the way it is, what it costs, and where it falls short.