AI & ML Quest 0 XP

How does a machine learn?

Your phone used AI dozens of times today: face unlock, maps, your feed, the spam folder. By the end of this quest you’ll know how it works, what it’s called and which AWS service builds it.

🗺️ 8 short levels 🎮 mini-games ⭐ earn XP and badges ⏱️ about 70 minutes 🤖 bonus: AI agents 🎬 video tutorial
Start level 1 See the big picture first ➕ More tutorials

This is your warm-up. After this quest, read the full Day 1 student guide: it will feel easy.

Your progress wheel: clear a level to light up a segment

1

Level 1 of 8, about 8 minutes

The AI family

Goal: explain AI, ML, deep learning and generative AI, and how they fit together.

Four words you hear every day. They are circles inside circles. Tap each circle.

Tap a circle to see its definition

Memory trick: always name the smallest circle. ChatGPT is AI, ML and deep learning, but its smallest circle is generative AI.

Why ML is different from normal programming

🧑‍💻 Traditional programming

Rules+DataAnswers

IF email contains “lottery” THEN spam.

🤖 Machine learning

Data+AnswersRules (a model)

Show 100,000 labelled emails. It learns what spam looks like.

Think of it like maths class. AI = the whole subject. ML = learning algebra from many solved examples. Deep learning = discovering advanced patterns yourself. Generative AI = now you can write new questions and a whole textbook.
🎯 Game: smallest circle
Which is the smallest circle this product belongs to?

🏁 Checkpoint

Which is always true?

What makes deep learning “deep”?

🎉 Level 1 cleared! +50 XP
2

Level 2 of 8, about 8 minutes

Data, and 4 ways to learn

Goal: sort data into types, and match a problem to a learning type.

Labelled or unlabelled?

🏷️

Labelled data

Every example has the answer. “This email is spam.” Slow and costly to create (often humans label it).

📦

Unlabelled data

Inputs only, no answers. A million transactions with no tags.

Structured, semi-structured or unstructured?

📊

Structured

Rows and columns. Excel, SQL tables, bank transactions.

🧾

Semi-structured

Tagged but flexible. JSON, XML, logs.

🖼️

Unstructured

No fixed format. Text, images, audio, video. Most of the world’s data.

🗂️ Game: sort the data
What type of data is this?

4 ways a machine learns. Tap to flip.

🧭 Game: which way to learn?
Pick the learning type for each problem.

🏁 Checkpoint

50,000 X-rays, each marked “pneumonia” or “normal”, are…

Grouping customers when you have no labels is…

🎉 Level 2 cleared! +50 XP
3

Level 3 of 8, about 12 minutes

How a model learns

Goal: describe training, the ML process, and overfitting.

🔢 Feature

An input. House size, location, rooms.

🎯 Label

The answer to predict. The price.

🧠 Model

The learned rule. price ≈ f(size…)

🏋️ Try it: train a model
The model is a line: price = w × size. It starts with a bad guess. Press “Train one step” and watch it predict, measure the error, and adjust w.
1. Predict2. Measure error (loss)3. Adjust w4. Repeat

The ML process is a loop

1. Business goal2. Frame the problem3. Collect data4. Clean data5. Features6. Train7. Evaluate8. Deploy9. Monitor
Fact: data work (steps 3 to 5) usually takes most of the time. Not training.

Never test on what you studied

📘 Training set, ~70–80%

Textbook exercises. The model learns here.

📝 Validation set, ~10–15%

Mock tests. Tune the settings.

🎓 Test set, ~10–15%

The real board exam. Used once, at the end.

🎚️ Try it: underfit, balanced, overfit
Slide the model complexity. Find the lowest error on new data.
Training error
New data error

Is 99% accuracy good? Not always.

🎯

Precision

When it says “fraud”, how often is it right? (Few false alarms.)

🕸️

Recall

Of all real frauds, how many did it catch? (Few misses.)

⚠️ If only 1% of transactions are fraud, a model that always says “not fraud” is 99% accurate and useless. Check precision and recall.

🏁 Checkpoint

99% on training data, 70% on new data. This is…

Which set gives the final, fair score?

🎉 Level 3 cleared! +50 XP
4

Level 4 of 8, about 6 minutes

Using the model: inference

Goal: choose between batch and real-time inference.

Inference = using a trained model on new data. Training is studying. Inference is sitting the exam.
📦

Batch inference

Many records at once, on a schedule. Nobody is waiting. Cheaper.
Example: nightly churn scores for all customers.
SageMaker AI: Batch Transform.

Real-time inference

One request, answer in milliseconds. Always-on endpoint.
Example: card fraud check while you pay.
SageMaker AI: real-time endpoint.

☁️

Serverless inference

No servers to manage. Scales to zero. Good for spiky or low traffic.

Asynchronous inference

Big inputs, long processing (e.g. long videos). Result delivered when ready.

Keyword trick 🔑 “immediately, low latency” → real-time. “nightly, large dataset” → batch. “intermittent traffic” → serverless. “large payload, long processing” → asynchronous.
⏱️ Game: pick the inference type
Read the situation. Which inference option fits best?

🏁 Checkpoint

During inference, the model’s weights…

A chatbot reply needs…

🎉 Level 4 cleared! +50 XP
5

Level 5 of 8, about 10 minutes

Deep learning: the digital brain

Goal: explain how a neural network learns, and name vision and language tasks.

🧠 Real brain🤖 Neural network
Dendrites receive signalsInputs (features)
Synapse strengthWeights
Fires only if the signal is strong enoughActivation function (e.g. ReLU)
Practice strengthens connectionsTraining adjusts the weights
🔌 Try it: build one neuron
A neuron multiplies each input by its weight, adds a bias, then ReLU keeps only positive results. Make it fire!
Input x₁
Weight w₁
Input x₂
Weight w₂
Bias b

How a network learns: 4 steps, repeated

➡️ Forward pass: predict📏 Loss: measure error⬅️ Backpropagation: find blame⛰️ Gradient descent: adjust weights
Think of walking downhill in fog. You can’t see the bottom (lowest error), but you feel the slope. Step downhill, feel again, repeat. The learning rate is your step size.

Right network for the job

🖼️ CNN

Images. Edges → shapes → objects.

📜 RNN / LSTM

Sequences, one step at a time. Older.

⚡ Transformer

Text and more. Self-attention sees all words at once. Powers LLMs.

👁️

Computer vision (CV)

Classification, object detection, segmentation, face recognition, OCR. AWS: Rekognition, Textract.

💬

Natural language processing (NLP)

Sentiment, entities, translation, speech-to-text, chatbots. AWS: Comprehend, Translate, Transcribe, Lex.

🔍 Game: which AI task?
Name the task being performed.

🏁 Checkpoint

In a neural network, synapse strength becomes…

Drawing boxes around every car and person in a photo is…

🎉 Level 5 cleared! +50 XP
6

Level 6 of 8, about 12 minutes

Generative AI: machines that create

Goal: explain foundation models, tokens, embeddings and diffusion.

Foundation model = a huge model pre-trained on broad data that you adapt to many tasks. An LLM is a foundation model for language.

The foundation model lifecycle

1. Data selection2. Pre-training3. Optimisation4. Evaluation5. Deployment6. Feedback
🔮 Try it: be an LLM
An LLM scores every possible next token, picks one, adds it, and repeats. Press the button to generate.
✂️ Try it: tokens
LLMs read tokens, not words. Rare words split into pieces. (Toy tokeniser: real ones learn their splits from data.) Rough rule: 1 token ≈ ¾ of an English word. Pricing is per token!
🗺️ Try it: embeddings, a map of meaning
Every word becomes a vector (a list of numbers). Similar meanings sit close together. Tap a word to see its nearest neighbours.
Tap a word

How AI creates images

🌫️ Try it: diffusion
Slide right: forward diffusion adds noise until only static is left (used in training). Slide left: reverse diffusion removes noise step by step. That is how the model creates an image.
CleanPure noise

t = 0

Multimodal models handle more than one type of data: text, images, audio, video. Upload a photo and ask a question about it.

🏁 Checkpoint

An LLM writes by…

Reverse diffusion…

🎉 Level 6 cleared! +50 XP
7

Level 7 of 8, about 10 minutes

Getting better answers

Goal: choose between prompt engineering, RAG, fine-tuning and RLHF.

🧪 Try it: prompt lab
Task: a leave email to your head of department. Switch on prompt ingredients and watch the output improve. Then play with temperature.
Temperature

The ladder: start at the bottom

TechniqueWhat changesCostUse when
1. 🧪 Prompt engineeringOnly your instructions💲Always first
2. 📚 RAGModel gets your documents at question time💲💲Private or changing facts, with sources
3. 🎓 Fine-tuningModel weights, with labelled examples💲💲💲Consistent style, format or skill
4. 📖 Continued pre-trainingWeights, with lots of domain text💲💲💲💲New domain language (legal, medical)
📝

Instruction fine-tuning

Train on “instruction → ideal answer” pairs. Turns a base model into an assistant.

🏅

RLHF

Humans rank answers → a reward model learns their taste → the model improves.

📚 Try it: follow a RAG question
RAG = open-book exam. Press “Next step” to follow the question through.
1. Ask“How many casual leaves do I get?”
2. EmbedQuestion → vector
3. RetrieveFind similar chunks in the vector DB
4. AugmentQuestion + HR policy text in the prompt
5. Generate“12 per year (HR Policy, p.4)”

On AWS: Amazon Bedrock Knowledge Bases does all five steps for you.

Fine-tuning vs RAG in one line
🎓 Fine-tuning = send the model back to school (changes how it behaves). 📚 RAG = hand it the right textbook page during the exam (changes what it knows today).
🧰 Game: pick the technique
Which technique fits best?

🏁 Checkpoint

Company policies change every month. The chatbot must use the latest ones. Use…

For consistent, factual answers, set temperature…

🎉 Level 7 cleared! +50 XP
8

Level 8 of 8, about 12 minutes

The AWS toolbox

Goal: pick the right AWS service, and know what drives the bill.

🍽️

AI services

Order from a restaurant: call an API, no ML skills. Rekognition, Polly, Comprehend…

👩‍🍳

ML platforms

A full kitchen: build your own. SageMaker AI, Bedrock.

🥕

Infrastructure

Raw ingredients: GPUs, Trainium (training), Inferentia (inference).

Memory hooks. Tap to flip.

Generative AI on AWS

🪨

Amazon Bedrock

Many foundation models (Nova, Claude, Llama…) through one API. No servers. Plus Knowledge Bases (RAG), Agents, Guardrails.

🚀

SageMaker JumpStart

Hub of pre-trained models inside SageMaker AI. Deploy and fine-tune on your own endpoint: more control.

💼

Amazon Q Business

Assistant for employees, connected to company data.

🧑‍💻

Amazon Q Developer

Helps developers and cloud engineers, e.g. in the AWS Console.

📢 Status updates
• Amazon Q Developer IDE plugins: no new sign-ups since 15 May 2026, support ends 30 April 2027. Replacement: Kiro. Q Developer in the AWS Console continues.
AWS DeepRacer: the console ended in December 2025. It lives on as an open-source solution you deploy in your own account.
Bedrock or SageMaker? “Least operational overhead, no infrastructure” → Bedrock. “Full control over the model and endpoint” → SageMaker AI / JumpStart.
🏁 Game: which AWS service?
Pick the service for each problem.
🧮 Try it: estimate an LLM bill
LLMs charge per input token and per output token. Change the numbers. Prices here are made up for practice: check the Amazon Bedrock pricing page for real ones.

Tokens per month:

Estimated cost per month:

🔤 Tokens

Input + output. Keep prompts lean.

📏 Model size

Smallest model that’s good enough.

📦 Batch

Much cheaper for non-urgent jobs.

🔒 Provisioned throughput

Reserved capacity for steady, heavy use.

💤 Idle endpoints

Real-time endpoints bill while running.

🎓 Customisation

Fine-tuning adds cost. Try prompts and RAG first.

🏁 Checkpoint

Speech → text is…

Call foundation models through one API with no servers to manage:

🎉 Level 8 cleared! +50 XP. Your wheel is complete!

More tutorials

Two extra pages. Your XP carries over.

🤖

Rise of the AI Agents: from talking to doing

Bonus level: agents, the think–act–observe loop, tools, safety and agents on AWS. About 12 minutes.

Open the bonus level
🎬

The Goldilocks Problem: how does an AI know when to stop learning?

Video tutorial on the bias-variance trade-off, with a scene-by-scene walkthrough and games. About 10 minutes.

Watch the video tutorial
🗺️

The big picture: mind map

Everything on one page. Click a branch to jump to its level.

🃏

Flashcards

Tap to flip. Mark each card. Keep going until the “review again” pile is empty.

Myth or fact?

Quick-fire. Decide fast.

👾

Boss battle

10 questions. 3 hearts. Lose a heart for each wrong answer. Can you beat the Overfit Monster?

Quest progress

AI & ML Explorer

📖 Next step: read the Day 1 student guide

You now know the map. In the guide, look for:

  • The 88 questions answered from basic to advanced, one module at a time
  • The ML process table with the SageMaker AI tool for each step
  • The comparison of prompt engineering, RAG and fine-tuning
  • The full AWS service table and the cost considerations
  • The Final challenge: aim for 10 out of 12