Skip to content

Learning path · AI Engineer

Become an AI engineer, one shipped project at a time

Build the AI inside products: prompts, retrieval, agents that use tools inside limits you set, and evals that catch a regression before users do. Start once you can write basic JavaScript or Python.

  • 11 courses
  • 162 lessons
  • Estimated 18–26 weeks · about 30 minutes a day
  • Included in Pro
  • In English, Uzbek, Korean, Russian, Spanish, French, German and Arabic
Start the AI Engineer path

Eligible first-time members get 7 free days. Cancel anytime.

AI Engineer

What you'll ship

Build and run a real AI product: it searches your own content, uses only the tools you approved, is tested the same way every time, has a plan for when it fails, and shows its speed and cost.

Path milestone: A support agent you can deploy: it searches your docs, uses tools safely, is tested, has a fallback, and stays inside a cost limit

Why it works

  1. Real engineering, plainly explained

    Tokens, context, structured outputs and retrieval, each in a short lesson with practice.

  2. Agents with limits

    Tool use, MCP and approval gates, so an agent only does what you allowed.

  3. Reliable after launch

    Evals, fallbacks and cost limits that keep an AI feature working once people use it.

Courses in this path, in order

  1. Prompting & LLM FoundationsPrompt clearly and get back structured output you can trust.16 lessons
  2. Retrieval-Augmented GenerationGive your LLM a memory it can trust.16 lessons
  3. Coding With AI AgentsBrief, steer, and ship the agent that writes your code.16 lessons
  4. Lead a Team of Coding AgentsSet the rules, review the work, and run several coding agents in parallel.18 lessons
  5. Secure Software BoundariesRepair isolated vulnerabilities and prove the safe behavior.9 lessons
  6. Build an AI AgentGive the model hands — tools, loops, and guardrails.16 lessons
  7. Evals: Make AI ReliableMeasure, trace, and trust your AI like a test suite.16 lessons
  8. Connect AI to Tools (MCP)MCP and safe tool use — give models real, guarded access.18 lessons
  9. Build MCP Servers & SkillsShip an MCP server and a skill that a real client actually loads.16 lessonsRequires Ultra
  10. System DesignTurn requirements into explicit architecture tradeoffs.9 lessons
  11. Developer Environment & DeliveryMove code from a clean checkout to a safe, observable release.12 lessons