AI Engineering Full Course - Build and Deploy a Real AI App
Dave Ebbelaar
A single project from start to finish: an AI agent that answers customer support emails. It reads a Gmail inbox, runs on a FastAPI backend with Pydantic AI and Claude, stores data in Supabase and searches a knowledge base for answers. Evals check the replies, a React admin dashboard shows them, and the system is deployed with Docker on a VPS. Some Python experience is assumed.
What you'll learn
- Reading and replying to emails with the Gmail API
- Building an agent pipeline with Pydantic AI and Claude
- Answering from a knowledge base stored in Supabase
- Writing evals for the agent's replies
- Deploying the system to a server with Docker
- An admin dashboard in React
