This course contains the use of artificial intelligence.

22 animated lessons · ~2 hours 43 minutes · zero agent experience needed. Go from 'what is an agent?' to shipping production-grade AI agents — and answering the interview questions that come with them.

You will build the systems real teams ship: the core agent loop, tool use and function calling, retrieval (RAG) with grounded answers and refusal behaviour, Model Context Protocol (MCP) servers built and tested end to end, human-in-the-loop approvals that survive failure, and multi-agent crews with budgets, traces and evals. Every number in this course comes from code that was actually run — eval scores, token costs, success rates — shown on screen, never invented.

Includes: 4 downloadable PDF books (code examples + practice exercises with solutions for every lecture), quizzes and coding exercises throughout, and a full interview track — agent loop, RAG, multi-agent design and a senior playbook — so you finish able to build agents AND talk about them like an engineer who has shipped them.

Every project in this course is built the way it would be built on a real team: requirements first, code you can run yourself, tests that prove it works, and a production checklist before anything ships. You will debug real failure modes — runaway loops, runaway costs, tools that fail, agents that refuse — and learn the patterns senior engineers use to keep agents safe, observable, and affordable in production.