Prompt Engineering Roadmap — 18 stations

This roadmap is interactive when JavaScript is enabled. Below is the full station list, grouped by track.

Foundation · Start here · ~4h

F1 What is a Prompt? — tokens, context windows, system/user/assistant roles, temperature basics. 30 min · Beginner
F2 Roles & Personas — system prompt design, tone & voice, persona stability. 30 min · Beginner
F3 CRISP Framework — Context · Role · Instruction · Specifics · Proof. 45 min · Beginner
F4 Bad vs Good Prompts — failure mode recognition, diagnosis. 30 min · Beginner
F5 Output Formatting — JSON mode, markdown, schemas, format consistency. 45 min · Intermediate
F6 Few-Shot Prompting — example selection, in-context learning, 0/1/few-shot. 60 min · Intermediate

Technique · Level up · ~5h

T1 Chain of Thought — step-by-step reasoning, zero-shot CoT. 45 min · Intermediate
T2 Self-Consistency — multiple sampling, majority voting, output aggregation. 45 min · Intermediate
T3 Tree of Thought — branching, evaluation, backtracking. 60 min · Advanced
T4 RAG Prompting — context injection, grounding, hallucination reduction. 60 min · Advanced
T5 Prompt Chaining — I/O contracts, state passing, pipeline design. 60 min · Advanced
T6 Meta-Prompting — self-critique, prompt generation, adversarial testing. 45 min · Advanced

Production · Ship it · ~6h

P1 Prompt Templates — placeholders, version control, reuse patterns. 45 min · Intermediate
P2 Token Optimization — compression, cost estimation, model selection. 30 min · Intermediate
P3 Safety & Guardrails — injection, refusal patterns, output validation. 60 min · Intermediate
P4 Evals & Testing — golden datasets, LLM-as-judge, regression. 90 min · Advanced
P5 Agent Prompting — ReAct, tool descriptions, loop termination. 90 min · Advanced
P6 Production Monitoring — logging, tracing, quality alerts. 60 min · Advanced

Roadmap · 2026 Edition

Prompt
Engineering.

18 stations. 3 tracks. Beginner to production-grade prompting — at your own pace.

Foundation

~4h 0/6

Technique

~5h 0/6

Production

~6h 0/6

0 of 18 stations · ~0h of ~15h

Begin the journey ↓

Lines —

Foundation

Technique

Production

Stations —

Not started

✓

Completed

The roadmap.

Three tracks. 18 stations. Click any node to open its detail. Mark complete as you go — your progress is saved locally.

Practice tools

Go deeper.

Interactive tools to practice what you've learned from the roadmap above.

Interactive tool

Build a CRISP prompt.

Fill in each component. Your prompt assembles live on the right.

Context — situation & background

Role — who the AI should be

Instructions — what to do

Specifics — format, length, constraints

Proof — examples of good output

Your prompt

Start typing in any field to see your prompt take shape…

Interactive lab

The Technique Lab.

Pick an experiment. See how each technique transforms a prompt — before and after.

Adding CRISP components to a vague prompt produces dramatically better output.

Before Vague prompt

Write a blog post about AI.

After CRISP prompt

Context: Our audience is non-technical founders who are curious about AI but intimidated by jargon.

Role: Act as a tech journalist who explains complex topics in plain English.

Instruction: Write a 600-word blog post introduction about how AI is changing content creation.

Specifics: Start with a hook anecdote. Use short paragraphs. Avoid technical terms — if you must use one, define it.

Proof: Tone example: "AI doesn't replace writers. It gives them a co-pilot that never sleeps."

AI response

Click "Simulate response" to see the difference.

Chain-of-thought forces the model to show its reasoning before answering — reducing errors on logic tasks.

Before Direct prompt

A train leaves City A at 60 mph. Another leaves City B (400 miles away) at 80 mph toward it. When do they meet?

After Chain-of-thought

A train leaves City A at 60 mph. Another leaves City B (400 miles away) at 80 mph toward it. When do they meet?

Think through this step by step. Show your reasoning clearly before giving the final answer.

AI response

Click "Simulate response" to see how CoT changes the output.

Without context, the model guesses. With injected document context, it answers from facts.

Before No context

What did customers say about our onboarding in Q4?

After RAG prompt

Answer using ONLY the context below. If the answer isn't there, say "Not found."

[CONTEXT]
Q4 Customer Feedback Summary:
- Revenue up 34% YoY to $4.2M
- Churn dropped from 8% → 5%
- Top complaint (38%): onboarding takes too long
- NPS: 72 (+8 pts vs Q3)

[QUESTION]
What did customers say about onboarding in Q4?

AI response

Click "Simulate response" to see how context grounds the response.

Keep reading.

The Prompting Handbook covers the Foundation track in depth — interactive, no code required.

Read the handbook →

PromptEngineering.

The roadmap.

Go deeper.

Build a CRISP prompt.

The Technique Lab.

Keep reading.

Prompt
Engineering.