The AI Engineer PathAI Engineer Path
Every module on the AI Engineer Path has a section here. Each one starts with a guide for running that module's project on your own machine, and grows as more written material for the module is added.
These pages are not the course. They exist so that the parts most likely to change, such as package versions, provider settings and model names, can be corrected in writing instead of by re-recording a lesson.
Modules
Intro to AI Engineering builds Gift Genie, first browser-only and then with an Express backend.
Deployment puts Dream Catcher online, with SQLite or PostgreSQL behind it.
Open-source Models runs open-source models three ways: through Hugging Face Inference, in the browser with Transformers.js, and on your computer with Ollama.
Embeddings and Vector Databases builds a chatbot on OpenAI embeddings and a Supabase vector store.
Intro to AI Agents fills the path's Agents slot. It has its own section, because it also runs as a standalone course.
Context Engineering manages what goes into a model's context window with system prompts and summarization, then practises it in a challenge on OpenRouter.
Vercel AI SDK builds a server-side customer-support agent that answers from support docs and calls tools.
Model Context Protocol builds and tests an MCP server in TypeScript.
Multimodality covers image generation and image understanding.
Already downloaded a project?
Go straight to Run an AI course project locally and pick your course from the list.