Academy / AI Engineering with Claude

AI Engineering with Claude

Prompting to production agents โ€” 7 levels, ~91 hours, 100% free.

This is the path I actually used. My 24-agent system that develops my own codebase came from these exact skills โ€” read the build notes at /notes/build-org/ when you reach Level 7.
0/7 levels ยท 0%
  1. How LLMs Actually Work (~6h) โ€” You can explain tokens, context windows, and temperature to a colleague without hand-waving.
  2. Prompt Engineering (~8h) โ€” You write prompts that behave the same on run 10 as on run 1 โ€” and you know how to test that.
  3. Claude API Fundamentals (~10h) โ€” You can call the Messages API from code: system prompts, streaming, vision, and honest error handling.
  4. Structured Outputs & Tool Use (~12h) โ€” Claude can call YOUR functions โ€” and you know how to make it decide correctly when to.
  5. RAG โ€” Retrieval Augmented Generation (~15h) โ€” You can make Claude answer from YOUR documents with citations โ€” and say 'I don't know' when it should.
  6. Agents & Orchestration (~20h) โ€” You can design multi-step agent systems โ€” and know when a workflow beats an agent.
  7. Production AI Systems (~20h) โ€” Your agent survives real-world failure: dead API keys, timeouts, cost spikes, and regressions.