PromptMarket

Roadmap

Learn how AI apps actually work.

Short lessons. Each one ends at a prompt you can copy.

What LLMs are good at

Start from a job the model can actually do.

  1. 01What LLMs are good atStart from a job a language model can do, not from a feature name.
  2. 02Structured dataTurn unstructured text into fields your program can use.
  3. 03ClassificationAssign an input to one label from a list you control.
  4. 04SummarizationCompress text while keeping the parts a specific reader needs.
  5. 05Question answeringAnswer a question from context you provide, and stop when that context is not enough.
  6. 06Agents and actionsLet a model request an action your application knows how to perform.

Prompting fundamentals

Instructions, context, examples, and the shape of the answer.

  1. 07Prompting fundamentalsA prompt is a job description: role, task, context, constraints, examples, and output shape.
  2. 17How to write AI agent promptsWrite an AI agent prompt with a clear objective, tool boundaries, examples, stopping conditions, and checks you can evaluate.

Application patterns

The usual ways an AI feature is assembled.

  1. 08Structured outputsRequire the model to produce data that matches a known structure.
  2. 09RAGFetch useful information first, then include it in the model's context.
  3. 10Tool callingGive the model a list of actions it can request.
  4. 11WorkflowsSplit one problem into several predictable model calls.
  5. 12Agentic loopsLet the model choose actions repeatedly until a stopping condition.

Building reliable AI

Treat model behavior as something you measure.

  1. 13AI engineering is experimentalThe first version will be wrong in specific ways. Keep those examples and change the system against them.
  2. 14EvalsAn eval is the AI equivalent of a unit or regression test.

Optimization

Add complexity only after the simpler version fails.

  1. 15Optimization ladderUse the simplest technique that solves the problem. Later rungs cost more.

Building AI products

A short loop from a useful problem to a better system.

  1. 16Building AI productsPick a useful problem, ship a small baseline, and let real failures decide the next change.