← Learning paths

8 lessons · take it step by step

Generative AI

Build reliable LLM applications: prompts, context, structured outputs, retrieval, tools and evaluation. Follow the eight lessons in order or start with the concept you need.

Start here

Follow the order if you are new, or jump straight to a topic you need.

Start the first lesson →

  1. LLM Application Fundamentals

    15–20 min

    Design an application around probabilistic output, finite context, latency and measurable quality.

  2. Prompt Engineering

    15–20 min

    Turn a vague request into a versioned task contract with examples and explicit failure behavior.

  3. Context Engineering

    15–20 min

    Select the right evidence, history and tool results within a bounded context budget.

  4. Structured Outputs

    15–20 min

    Validate a machine-readable response before using it in application logic.

  5. Embeddings & Vector Search

    15–20 min

    Understand similarity ranking, embedding compatibility and retrieval trade-offs.

  6. RAG Fundamentals

    15–20 min

    Build a source-aware retrieval pipeline with explicit abstention and access filtering.

  7. Tool Calling

    15–20 min

    Treat a model tool request as untrusted input and execute it through an application policy.

  8. LLM Evaluation

    15–20 min

    Create a repeatable quality gate for task success, evidence use, safety, latency and cost.