Prompt Engineering
Stop guessing at prompts. Learn how LLMs read your words, then how to write prompts that work — from anatomy and system prompts to tool calling, chaining, evals, and safety.
Prompt Engineering: Zero to Mastery
A complete, beginner-friendly curriculum that takes you from “what is a prompt?” all the way to building, evaluating, and shipping production-grade LLM systems.
Read the modules in order. Each one builds on the last. Every module has explanations in plain language, copy-paste examples, “try it yourself” exercises, and common mistakes to avoid.
Who this is for
- You have never written a prompt before, OR
- You write prompts but they feel like guesswork, AND
- You want to become genuinely expert — the person on the team who others ask for help.
No coding background is required to start. Modules 6 and later show code (Python), but every concept is explained from scratch and you can understand 80% of the value without writing a single line of code.
How to use this curriculum
- Read actively. Don’t just read the example prompts — open a chatbot (Claude, ChatGPT, Gemini, etc.) in another window and actually run them.
- Do the exercises. Each module ends with exercises. Skipping them is the #1 reason people stay stuck at “beginner.”
- Keep a prompt journal. A simple text file where you save prompts that worked and prompts that failed, with a note on why. This becomes your personal playbook.
- Don’t rush. Mastery comes from reps, not from reading fast.
The roadmap (the modules)
| # | Module | What you’ll learn | Level |
|---|---|---|---|
| 01 | Foundations: How LLMs Actually Work | Tokens, context windows, temperature, why models “hallucinate” | Beginner |
| 02 | Anatomy of a Great Prompt | The building blocks: role, task, context, constraints, examples, format | Beginner |
| 03 | System Prompts | Setting persistent behavior, persona, rules, and guardrails | Beginner→Intermediate |
| 04 | Core Prompting Techniques | Zero-shot, few-shot, chain-of-thought, self-consistency, decomposition | Intermediate |
| 05 | Structured Output (JSON & Schemas) | Forcing reliable, machine-readable output | Intermediate |
| 06 | Tool Calling / Function Calling | Letting the model use calculators, search, APIs, databases | Intermediate→Advanced |
| 07 | Prompt Chaining & Workflows | Breaking big tasks into reliable multi-step pipelines | Advanced |
| 08 | Advanced Techniques: RAG, ReAct & Agents | Retrieval, reasoning+acting loops, autonomous agents | Advanced |
| 09 | Evaluating Prompts (Evals) | Measuring quality, building test sets, LLM-as-judge, regression testing | Advanced |
| 10 | Safety, Security & Prompt Injection | Jailbreaks, injection attacks, guardrails, responsible use | Advanced |
| 11 | Production Best Practices | Cost, latency, caching, versioning, monitoring, model selection | Advanced |
| 12 | Capstone Projects & Continued Learning | Build real projects, a study plan, and where to go next | Mastery |
A 6-week study plan (optional)
- Week 1: Modules 01–02. Run every example. Start your prompt journal.
- Week 2: Modules 03–04. Rewrite 5 of your old prompts using new techniques.
- Week 3: Module 05–06. Build one structured-output prompt and one tool-calling demo.
- Week 4: Modules 07–08. Build a small 3-step chain (e.g., research → outline → draft).
- Week 5: Modules 09–10. Create an eval set for one of your prompts; try to “break” it.
- Week 6: Modules 11–12. Polish one project end-to-end and write it up.
A note on terminology
This field has a lot of jargon. Every term is defined the first time it appears, and there’s a glossary at the end of Module 01. When in doubt, search for the term in that glossary.
Let’s begin. Open Module 01: Foundations.