Prompt Engineering
Get better results from any model — by design, not by luck. Learn clear prompting, few-shot examples, system roles, advanced techniques and how to fix bad output, and remember it with spaced repetition.
- practice cards
- 90
- practice cards
- per day
- ~10 min
- per day
- level
- Beginner → Intermediate
- level
- modules
- 5
- modules
What is prompt engineering?
Prompt engineering is the craft of getting reliable, useful output from a language model by how you ask. The same model can give a vague, wrong answer or a precise, structured one depending entirely on the prompt — so a few repeatable techniques go a long way.
This track covers the toolkit: writing clear, specific instructions, steering behaviour with system prompts and roles, teaching the format with few-shot examples, advanced moves like chain-of-thought, and how to diagnose and fix output that misses the mark.
It uses spaced repetition so the techniques become second nature — and it pairs naturally with AI & LLM Fundamentals (why prompts work) and Building with LLMs (putting them into apps).
5 modules, seed to bloom
Each module is a set of practice cards — 90 in total. Answer, review, and watch your knowledge grow from seed to full bloom.
Prompting Basics
The foundation of good prompts — clear instructions, context, output format, and length
18 cardsExamples & Few-shot
Teaching by showing — zero-, one-, and few-shot prompting and in-context learning
18 cardsRoles & System Prompts
Steering behavior — system prompts, personas, role prompting, and instruction priority
18 cardsAdvanced Techniques
Getting more from prompts — chain-of-thought, task decomposition, prompt chaining, and self-consistency
18 cardsPitfalls & Fixing Outputs
Spotting and fixing bad output — common prompt mistakes, hallucinations, and iterating effectively
18 cardsSample questions
A taste of the real cards. Pick an answer, then reveal the explanation.
What is "zero-shot" prompting?
- AAsking with no examples — you give only the instruction and let the model answer
- BAsking the same thing twice — you send a prompt again to confirm the first reply
- CAsking with the temperature off — you disable randomness so output never varies
- DAsking without any context — you strip out background to save on the token usage
What is a "system prompt"?
- ATop-level instructions — they set the model's behavior for the whole conversation
- BAn automatic error report — a message logged when the model crashes mid-reply
- CThe system's hardware specs — the server details the model runs its work upon
- DA status update from the app — a notice telling you the service is online now
What is "chain-of-thought" prompting?
- AAsking the model to reason step by step — working through its logic before answering
- BAsking the model to answer instantly — skipping any reasoning to save on the time
- CLinking many models in a chain — each one refines the previous model's own output
- DChaining user messages together — merging a whole chat into one single request
What is a common cause of vague, off-target answers?
- AAn under-specified prompt — the model fills the gaps with its own assumptions
- BA model that is too large — bigger models tend to wander off the question
- CToo low a temperature — precise settings make answers drift from the topic
- DA missing internet link — without live access the model loses focus entirely
Learn it once, keep it for good
Answer a question
Each card is one practical concept with multiple options. Pick what you think is right.
Get the full answer
See the correct option plus a clear explanation, and a link to deeper docs when one is available.
Review at the right time
A spaced-repetition engine (SM-2 or FSRS) resurfaces each card just before you would forget it.
Why prompt engineering is worth your time
Works with every model
These techniques are model-agnostic — they make you better with whatever assistant or API you use.
Repeatable, not lucky
Stop tweaking prompts at random. Understand why a prompt works and reproduce the result on purpose.
Fix bad output fast
Learn to diagnose vague, off-format or wrong answers and correct them with a targeted change.
A daily productivity multiplier
Better prompts mean less back-and-forth on the tasks you already do with AI every day.
Common questions
Do I need to understand how LLMs work first? +
It helps but is not required. This track is practical and self-contained; for the why behind the techniques, the AI & LLM Fundamentals track pairs well with it.
How long does it take? +
About 10 minutes a day. Spaced repetition means short, frequent sessions beat long cramming, so the techniques stick.
Is it free? +
Yes, completely free. No registration or credit card is required, and all your progress is stored locally in your browser.
Is this tied to one specific AI tool? +
No. The techniques are model-agnostic and apply to any chat assistant or LLM API you work with.
Ready to write better prompts?
Plant your first seed today. Ten minutes a day is all it takes to grow prompting skills that work on purpose.