This course is about working with agentic AI on your own tasks and putting it to work inside an organization. It opens with what a large language model actually does, how the current models differ, and how to get dependable results from them on data and documents. It then turns to the machinery underneath — the tools and harnesses an agent runs on, what a sandbox does and does not contain, and the instructions that persist from one session to the next — and to customizing that machinery with skills and plugins you write and distribute. The later weeks build and evaluate AI-enabled software, ground it in your own documents through retrieval and fine-tuning, and end with the work that has to stay in-house: running models on your own hardware and de-identifying data before it goes anywhere else.