How this platform works
Learn how lessons, labs, cost labels, exercise tiers and the AI assistant fit together so you can study efficiently and safely.
Assumes: .NET 10 LTS ยท AWS SDK for .NET v4 ยท AWS CLI v2 ยท Region us-east-1. Verified against official sources on 2026-11-17.
Learning objectives
By the end of this lesson you can:
- Describe the path from module to lesson to lab to project, and pick the right next step.
- Read the page labels (
difficulty,minutes,cost,verifiedOn,versions) and decide whether you can do a hands-on page today. - Choose the right exercise tier (๐ข Guided, ๐ก Challenge, ๐ด Live problem) for the time you have.
- Use the AI assistant productively, and list what you must never paste into it.
- Re-check a lesson whose
verifiedOndate is old.
Prerequisites
- None. This is the first lesson. Start at the module overview if you have not read it.
Why this matters in banking
In a bank, the cost of a learning mistake is rarely the bill โ it is the incident. A developer who pastes a production connection string into a public chatbot, or who "just tries something" in a shared account, can trigger a data-protection breach report under UK GDPR, a security investigation and an uncomfortable conversation with the regulator. The habits in this lesson (separate sandbox, cost labels, no secrets in prompts, checking dates) are the same habits a bank expects from an engineer on day one.
The idea in plain language
This platform is a structured course plus a practice workshop plus a study partner:
- Lessons explain one idea at a time, always starting from what you already know in .NET and Azure.
- Labs are hands-on projects in the
labs/folder of the repository. Each one builds and passes its tests before it is published. - Projects and capstones combine many lessons into portfolio work you can show an employer.
- The AI assistant answers questions about the lesson you are on, helps you debug, and quizzes you.
Analogy: It works like a bank's graduate programme. Lessons are the classroom sessions, labs are supervised shifts in a training branch, capstones are your first real project, and the AI assistant is a senior colleague you can ask questions โ one who is helpful but occasionally wrong, so you check what they say.
What problem this structure solves
AWS has more than 200 services, and most tutorials either drown you in theory or skip the parts banks care about (identity, audit, cost, compliance). This structure solves three problems:
- Order. Modules have prerequisites. Module 0 makes your environment safe; Module 1 explains fundamentals; later modules assume both.
- Cost surprise. Every hands-on page states its cost up front, so you never discover a charge after the fact.
- Staleness. AWS changes weekly. Every factual page shows when it was last checked and against which versions.
Limitations to be honest about: we verify facts against official AWS documentation on the date shown, but AWS can change prices, console layouts and quotas at any time. The AI assistant can make mistakes. When the stakes are real, the official documentation and your organisation's policies win.
Microsoft / Azure comparison
| You know (Microsoft) | Here | What's different |
|---|---|---|
| Microsoft Learn modules and learning paths | Modules โ lessons โ labs โ capstones | Every lesson bridges from .NET/Azure and ends with cleanup |
| Azure free account credits | cost label on each page | You see the cost before you start, and a free alternative is offered for paid labs |
| GitHub Copilot Chat in Visual Studio | The platform's AI assistant | The assistant receives the lesson as context; treat it as a public service for data-protection purposes |
| Microsoft Learn "last updated" date | verifiedOn + versions | We record the date facts were checked, not just edited |
Architecture
How the parts connect:
Security-boundary view of the AI assistant: the assistant runs on this platform's backend and sends your question plus lesson text to an external large-language-model provider. Anything you type crosses that boundary. Assume it may be logged by the provider.
Step-by-step: how to study a lesson
Check the labels at the top
Read difficulty, minutes and cost. If the cost is credits or paid, confirm your cost guardrails are active before you start.
Read the plain-language section and the Azure comparison first
These two sections give you the mental model. If they do not make sense, ask the assistant to explain them differently before you go deeper.
Do the hands-on part in your sandbox account
Use the sandbox profile (learn-sandbox), tag what you create, and keep a terminal open with your cleanup commands ready.
Run cleanup immediately
Do not leave cleanup until tomorrow. Follow the lesson's "Cost and cleanup" section, then the teardown checklist.
Prove you learned it
Answer the knowledge check without looking, then try at least the ๐ข exercise. Tick the completion checklist and press Mark as complete.
The labels on every page
| Label | Meaning | What you do with it |
|---|---|---|
difficulty | beginner ยท intermediate ยท advanced | Skip ahead only if the prerequisites are solid |
minutes | Realistic study time including hands-on | Block the time; never stop half-way through a deployment |
cost: free | Runs locally or within always-free allowances | Go ahead |
cost: credits | Small spend, under $1 if you tear down the same day | Check budgets are active; start early in the day so you can finish and clean up |
cost: paid | Noticeable cost (for example a NAT Gateway or database left running) | Optional. A free, design-only alternative is always provided |
verifiedOn | Date facts were checked against official AWS sources | If older than 6 months, re-check prices and console steps |
versions | Version assumptions (.NET 10 LTS, SDK v4, CLI v2, Region) | If your versions differ, expect small differences |
โ VERIFY | A fact we could not confirm from an official source | Check it yourself before relying on it |
The three exercise tiers
| Tier | What it is | Typical time | When to use it |
|---|---|---|---|
| ๐ข Guided | Step-by-step task that repeats the lesson in a slightly different setting | 10โ30 min | Always do this one |
| ๐ก Challenge | A variation with a goal but no steps | 30โ60 min | When you want the knowledge to stick |
| ๐ด Live problem | An ambiguous, ticket-style situation like real work ("the bill doubled overnight โ find out why") | 30โ90 min | Before interviews and when you feel confident |
Solutions are in collapsible sections. Try first; open the solution only to compare, and read the why explanation even if you got it right.
AWS Console, CLI and IaC instructions
Not applicable for this lesson โ it describes how to study, not an AWS service. Hands-on lessons include all three.
Complete code
Not applicable โ the first runnable code is in the verify-setup lab.
Security considerations: using the AI assistant safely
The assistant is useful for explaining concepts, reading error messages and reviewing your own sample code. Treat it like any external service your bank has not approved for confidential data.
Never paste:
- Access keys, secret keys, session tokens, passwords, connection strings, private keys or certificates. If you paste one by mistake, treat it as leaked: rotate or delete it immediately.
- Real customer data of any kind โ names, account numbers, IBANs, card numbers, addresses, transaction histories. Under UK GDPR and GDPR, sending personal data to an unapproved processor can itself be a reportable incident.
- Your employer's code, architecture diagrams, account IDs, internal hostnames or incident details.
- The contents of
%USERPROFILE%\.aws\credentials,%USERPROFILE%\.aws\sso\cacheor%USERPROFILE%\.aws\login\cache.
Safe to paste: lab code from this platform, error messages with account IDs and ARNs replaced by placeholders such as 111122223333, and fictional data generated by the course's data generators.
โ Learning shortcut โ not for production
In this course you may paste sanitised error messages and sample code into the assistant. In a bank, follow your organisation's approved-AI policy: many banks allow only internally hosted or contractually approved assistants, and some forbid pasting any source code. Ask your information security team before using any AI tool on work material.
Also remember: the assistant can be wrong with confidence. Check commands against the official documentation before running anything that creates, changes or deletes resources, and never run a command you do not understand in an account that matters.
Banking use case
A fictional example: Priya, a developer at "Harbour Bank" (fictional), is learning AWS for a payments migration. She studies on her personal laptop in her own sandbox account. When her CDK deployment fails, she copies the error into the assistant after replacing the account ID with 111122223333 and removing the ARN of her role. The assistant suggests the error comes from a Region-restricting SCP. She confirms it in the AWS documentation, fixes her stack to use us-east-1, deploys, then runs cdk destroy and checks Tag Editor before closing the laptop. She never uses Harbour Bank data or accounts for study.
Common mistakes
- Studying in a work account โ it feels convenient โ it breaks policy and puts real data at risk. Use your own sandbox.
- Skipping cleanup "until later" โ you plan to continue tomorrow โ hourly resources bill all night. Tear down every session; redeploying with IaC takes minutes.
- Trusting the assistant blindly โ its answers sound authoritative โ AWS changes and models hallucinate. Verify against official docs.
- Ignoring
verifiedOnโ prices and console menus move โ re-check pages older than six months. - Pasting secrets "just once" โ debugging pressure โ treat any pasted secret as compromised and rotate it.
Troubleshooting
| Symptom | Likely cause | Fix |
|---|---|---|
| A console step does not match the lesson | AWS changed the console after verifiedOn | Use the console search bar for the feature name; check the linked official doc |
| A price differs from the lesson | Price change or a different Region | Check the official pricing page for us-east-1; prices vary by Region |
| The assistant contradicts the lesson | Either may be out of date | The official AWS documentation decides; report the discrepancy |
| The Mark as complete button is missing | You are on a module or reference page | Only lessons, labs and projects have it |
Cost and cleanup
This lesson costs nothing and creates no resources, so there is nothing to clean up.
Exercises
- ๐ข Guided: Open three lessons in later modules. For each, write down its
costlabel andverifiedOndate, and decide whether you could do it today with your current guardrails. - ๐ก Challenge: Write a one-paragraph "personal AI-use policy" for your studies: what you will and will not paste, and what you will do if you paste something by mistake.
- ๐ด Live problem: A colleague says, "I pasted our test environment's database connection string into a chatbot to debug a timeout. It's only test, so it's fine, right?" Decide what you would advise and who should be told.
Assignment
Create a private Git repository called aws-learning-journal. Add a README.md with your study schedule, your AI-use policy from the challenge, and a table you will fill in as you go: lesson, date completed, cost incurred, cleanup confirmed (yes/no).
Solutions
There is no single answer. The point is the decision rule: free pages can be done any time; credits pages only when budgets and alarms are active and you have time to clean up the same day; paid pages are optional and have a free design-only path. If any verifiedOn is older than six months, plan extra time to re-check prices. Why: reading the labels first turns cost from a surprise into a choice.
A good policy names concrete items: "I never paste keys, tokens, passwords, connection strings, real personal data, employer code or account IDs. I replace IDs with 111122223333. If I paste a secret, I revoke or rotate it within the hour and note it in my journal." Why: a written rule removes in-the-moment judgement, which is exactly when mistakes happen.
It is not fine. A connection string is a credential; "test" environments in banks often contain copies of real data or share networks with production. Advise: (1) rotate the password immediately, (2) check whether the test database holds real or masked customer data, (3) report it through the bank's security incident process โ the security and data-protection teams decide whether it is reportable, not the developer. Why: under GDPR / UK GDPR the controller must assess personal-data breaches quickly; hiding it is worse than the mistake.
Knowledge check
Answer: It promises the lab should cost less than $1 if you tear it down the same day. It does not promise zero cost, and it does not protect you if you forget cleanup โ hourly resources keep billing until deleted.
Answer: (b). (a) is a credential, (c) is personal data, (d) is your employer's confidential material.
Answer: Check the current prices on the official AWS pricing page for your Region. The page's facts were correct on that date but may have changed since.
Interview questions
- "How do you keep up with AWS changes?" โ I note the date and version of anything I learn, re-check official docs and What's New announcements when I use a feature again, and treat blog posts and AI answers as leads, not facts.
- "What's your approach to using AI assistants with company code?" โ I follow the approved-tools policy, never paste secrets or personal data, sanitise identifiers, and verify every suggestion before running it.
Certification practice
A developer pastes a customer's account number into a third-party chatbot while debugging an application hosted on AWS. Under the AWS shared responsibility model, who is responsible for this data-handling decision?
A. AWS, because the application runs on AWS B. The customer (the bank), because protecting its data and controlling who it is shared with is security in the cloud C. The chatbot provider, because it received the data D. Nobody, because the data never left AWS
Answer: B. AWS secures the cloud infrastructure; the customer is responsible for its data, identities and how its staff handle information. A is wrong because AWS does not control what your staff do with data. C is wrong because the bank chose to send it. D is wrong because pasting into an external chatbot does send it outside.
Completion checklist
- I can explain the difference between
free,creditsandpaid - I can name the three exercise tiers and when to use each
- I have written my personal AI-use policy
- I know what to do if I paste a secret by mistake
Key takeaways
- Modules have an order; finish Module 0 before deploying anything.
- Read the
costlabel before every hands-on page;creditsmeans "under $1 if torn down the same day". - Every exercise has three tiers โ always do ๐ข, push yourself with ๐ก and ๐ด.
- The AI assistant is a helpful colleague, not an authority, and never a place for secrets or customer data.
verifiedOntells you how fresh a page's facts are; re-check anything older than six months.
Further reading
Independent learning resource โ not affiliated with, endorsed or sponsored by Amazon Web Services. AWS and related marks are trademarks of Amazon.com, Inc. or its affiliates. All banking data in examples is fictional. Sample architectures are for learning and are not automatically compliant with GDPR, PCI DSS or any banking regulation.
Module 0: Orientation & a Safe Learning Environment
Set up a locked-down, cost-guarded AWS learning environment and a Windows toolchain before you build anything.
Secure your AWS account
Lock down the root user, sign in through IAM Identity Center, isolate labs in a sandbox account and restrict Regions with a service control policy.