Skip to content
skillbld
All courses

Learn how large language models work and how to build real apps with them: prompting, APIs, structured output, tools, retrieval-augmented generation (RAG) and agents. Visual explanations, hands-on Python you can run in the lesson, and a quick knowledge check every day.

  • 14 days
  • ~35 min / day
  • Intermediate
  • Free account required
0/14

Ready when you are

Saved on this device. No account needed.

Start Day 1

The big picture

Everything you'll cover, and how it fits together.

Mind map · click a dot to fold, drag to pan

Day by day

Go at your own pace. Your progress is remembered on this device.

  1. 1
    Day 1Up next 35 min

    What Is an LLM?

    Tokens, next-token prediction, sampling and context windows.

    TokensNext-token predictionContext window
  2. 2
    Day 2 35 min

    Prompting Fundamentals

    Roles, clear instructions, examples and structure.

    System promptsFew-shot examplesXML tags
  3. 3
    Day 3Coming soon 35 min

    Calling an LLM API

    Your first API call, API keys, parameters and costs.

    SDKsMessagesTokens & cost
  4. 4
    Day 4Coming soon 35 min

    Conversations & Streaming

    Keep chat history and show answers as they're written.

    Multi-turnStreamingStop reasons
  5. 5
    Day 5Coming soon 35 min

    Structured Output

    Get reliable JSON that your code can use.

    JSONSchemasValidation
  6. 6
    Day 6Coming soon 35 min

    Tool Use

    Let the model call your functions.

    Tool definitionsTool callsResults
  7. 7
    Day 7Coming soon 35 min

    Embeddings & Vector Search

    Turn text into vectors and search by meaning.

    EmbeddingsCosine similarityVector databases
  8. 8
    Day 8Coming soon 35 min

    RAG: Retrieval

    Chunk, index and retrieve your own documents.

    ChunkingIndexingTop-k search
  9. 9
    Day 9Coming soon 35 min

    RAG: Generation

    Ground answers in sources and cite them.

    GroundingCitationsHallucinations
  10. 10
    Day 10Coming soon 35 min

    Agents

    The agent loop: think, act, observe, repeat.

    Agent loopPlanningStopping
  11. 11
    Day 11Coming soon 35 min

    Multi-Step Agents & MCP

    Connect agents to tools and data with the Model Context Protocol.

    MCPSub-agentsMemory
  12. 12
    Day 12Coming soon 35 min

    Evaluating LLM Apps

    Measure quality before you ship changes.

    Test setsGradersLLM-as-judge
  13. 13
    Day 13Coming soon 35 min

    Safety, Cost & Production

    Prompt injection, guardrails, caching and monitoring.

    Prompt injectionCachingMonitoring
  14. 14
    Day 14Coming soon 35 min

    Capstone: A Docs Assistant

    Combine RAG, tools and an agent loop into one app.

    DesignBuildEvaluate