Interactive learning
Learn by making the engineering decision yourself.
16 lessons across 3 domains, on the decisions that are easy to get wrong and expensive to get wrong twice. Each one puts you in the situation, makes you choose, and then works out what your choice actually does.
Free, no sign-up
No account, no email, no trial.
Runs in your browser
The lessons compute locally; nothing you choose is sent anywhere.
Deterministic
The same choice always produces the same result. Nothing is graded by chance.
Reliability & Integration
6 lessons
What to do when a call fails, times out, or may have already happened. Six lessons on the decisions that keep an integration honest under failure.
Retry + Dead Letter Queue
Choose how an order delivery responds when the warehouse API is unavailable, then inspect each attempt and its outcome.
Retry · Fixed Backoff · Dead Letter Queue · Correlation ID
Start lab →Retry Decisions
What should you do with this failure?
The same HTTP status can require different actions. Learn to decide from the failure context — not from a cheat sheet.
Retry · Retry-After · Fail fast · Dead Letter Queue
Start lab →Timeout
Did it fail, or didn't it?
A timeout means the caller stopped waiting, not that the work failed. Learn what you actually know when the clock runs out, and what to do about it.
Timeout · Unknown outcome · Status check · Idempotency key
Start lab →Idempotency
Retry without doing it twice
The same request can arrive more than once. Learn how an idempotency key lets the receiver create the business effect only once, and why a correlation ID cannot do that job.
Idempotency key · Correlation ID · Safe replay · Conflict
Start lab →Circuit Breaker
When should nobody call at all?
A retry asks whether to send this request again. A breaker asks whether anyone should be calling that dependency right now. Learn what its state means and how it gets back to normal.
Circuit breaker · Fail fast · Cool-down · Half-open probe
Start lab →API Rate Limiting
How fast may we send?
HTTP 429 is not a fault report: it is the API telling you that you are sending too much. Learn to read your allowance, respect the signal, and pace work through a window.
Rate limit · HTTP 429 · Retry-After · Pacing
Start lab →Quantum
5 lessons
What a quantum computer can and cannot do, and what that means for systems you run today. Five lessons that grade the honest answer, including refusing to claim more than the evidence supports.
Superposition and Measurement
Is it 0 and 1 at the same time?
A qubit in superposition is not a coin already showing a face under the lid. The difference is measurable: ask it a different question and the coin gives noise where the qubit gives certainty.
Qubit · Superposition · Measurement basis · Pedagogical simulation
Start lab →No-cloning
Why can't you take a backup?
The obvious copier — measure it, then make two of what you saw — copies |0⟩ and |1⟩ perfectly and quietly destroys anything else. No better device exists: copying an arbitrary unknown state is ruled out by a theorem.
No-cloning theorem · Measure and prepare · Quantum key distribution · Pedagogical simulation
Start lab →Entanglement
Is it just correlation?
Two gloves in two boxes are correlated: open one and you know the other. An entangled pair agrees in every question you can ask it, and still carries no message. Telling those apart is the lesson.
Entanglement · Shared cause · No-signalling · Bell test
Start lab →Quantum Advantage
What would actually go faster?
A quantum computer is not a faster computer. For two kinds of problem the advantage is enormous, for one it is quadratic and usually swallowed by overhead, for most there is none at all — and for some, nobody knows yet.
Quantum advantage · Shor · Grover · Error-correction overhead
Start lab →Harvest Now, Decrypt Later
Which of these is already late?
Traffic captured today can be opened whenever the capability arrives, so the clock started when the data was sent. Three numbers decide each asset — and two of the seven here are not exposed at all.
Harvest now, decrypt later · Mosca's inequality · Crypto-agility · Post-quantum migration
Start lab →Agentic AI
5 lessons
What an agent is allowed to do, what it actually knows, and when a human has to decide. Five lessons on giving an agent real tools without giving it the whole estate.
Tool Authority
What may the agent actually do?
An agent is exactly as dangerous as the tools it holds. Each request has three numbers — the authority asked for, the authority needed, and how many records one call could touch — and the gap between them is the accident waiting to happen.
Least privilege · Blast radius · Irreversible actions · Tool surface
Start lab →Belief and State
How does the agent know?
An agent's working state is a model of the world, not the world. Three things have to stay apart: what it believes, what the system knows, and what actually happened — and only the second one is authoritative.
Verification · Stale reads · Agent belief · Evidence
Start lab →Agent Retries
When should the loop try again?
A person retries once and thinks about it. A loop retries as fast as the network allows, several times, without pausing — so the retry policy is no longer advice, it is behaviour.
Idempotency keys · Retry budgets · Agent loops · Escalation
Start lab →Memory and Context
What should the agent keep?
A memory feature is a database with a friendly name. Four different decisions hide inside the word "remember", and two of them are refusals.
Agent memory · Context · Data retention · Stale state
Start lab →Human Approval and Irreversible Actions
Which of these may it just do?
Approving everything and approving nothing fail the same way. What the rules already say, whether it can be undone, how wide it reaches and how often it happens decide it.
Human in the loop · Irreversible actions · Approval fatigue · Dry runs
Start lab →How a lesson works
Step 1
You are given a situation, not an article
A real failure, a real ambiguity, a real trade-off — with the evidence an engineer would actually have at that moment.
Step 2
You decide
Every lesson has a tempting wrong answer that looks right for a reason. Choosing it is the fastest way to find the edge of your mental model.
Step 3
The result is worked out, not asserted
Your choice runs against the lesson's model and you see the consequence, the evidence behind it, and why the other answers behave differently. Reset and try the other one.