Phase 04 · Modern data platforms · cloud
Cloud Platforms for Data Engineers
- lessons
- 14
- modules
- 4
- stages
- 4
- of learning
- 2h 58m
Concept Traces in this track
Playable walkthroughs: watch the system move, predict the next step, stamp a memory seal, then practice. Completing a Trace counts toward readiness.
Why this track exists
Everything you have built so far runs on your laptop. Your laptop sleeps, has one disk, and is not a place a company will keep its customer data. So the work moves to rented computers, which introduces three new problems that have nothing to do with data: who is allowed to touch what, where files live now that there is no C drive, and who pays for the query you just ran. Cloud platforms are the answer to those, and the reason a badly written query can cost real money.
You will spend more time on permissions than you expect. A large share of real cloud work is IAM, storage layout, and cost, not clever architecture. Knowing the shape of the services and the vocabulary across providers is what lets you be useful in the first week.
What you need before starting
Basic pipeline experience
You should have written something that reads and writes data. The cloud versions make far more sense with that reference point.
No cloud account needed
This track walks through provider-neutral cloud primitives first, with AWS, Azure, and GCP implementations at equal depth. The exercises are simulated with zero bill.
The roadmap
4 stages, in the order they build on each other. Each stage lists the modules and lessons it covers, and what you should be able to do by the end of it.
01Provider-neutral foundations
Cut through the marketing. It is someone else's computers, rented by the second, with an API in front of them. Everything else follows from that.
What is the CloudFree
0/4
What cloud computing is, the main service types, data lakes versus warehouses, and security basics.
By the end of this stage
You can explain what you are actually renting, and why it is billed the way it is across compute, storage, and networking.
02The mental model: identity and storage
Two things underpin every cloud data system: who is allowed to do what, and where the bytes live. Compare S3, ADLS Gen2, and GCS alongside AWS IAM, Entra ID RBAC, and GCP IAM.
Cloud Mental ModelFree preview
0/3
Which cloud, who can do what, and where bronze files actually land.
- Why the cloud, and which one12m
- IAM: who can do whatFree14m
- Object storage as the bronze landing zone14m
Module checkpoint
2 conceptual questions to verify mastery
By the end of this stage
You can reason about IAM roles, managed identities, and object storage layout across AWS, Azure, and GCP, and explain why folder prefixes are a design decision rather than a detail.
Where people get stuck
IAM is boring right up until it is the reason your job fails at 3 a.m. or the reason data leaked. Learn it properly now.
03Managed data services across clouds
Warehouses (Redshift, Synapse, BigQuery), managed Spark (EMR, Synapse Spark, Dataproc), and managed schedulers (MWAA, Data Factory, Composer) each replace something you built by hand.
Managed Data Services
0/4
Warehouses, orchestrators, Spark clusters, and the function that fires when a file lands.
Module checkpoint
1 conceptual questions to verify mastery
By the end of this stage
You can design a data pipeline architecture and map every component to its equivalent on AWS, Azure, and GCP with equal depth.
04Cost governance, security, and shipping it
In the cloud, a design decision is a line item. Scanning a whole table because you forgot a partition filter has a price, and someone will ask about it.
Cost, Security, and Shipping It
0/3
Read a bill, know why a job cannot reach a database, then land ingest on a real free-tier project.
- Reading a cloud bill12m
- Networking a data engineer actually needs12m
- Capstone: deploy ingest to the cloud16m
Module checkpoint
1 conceptual questions to verify mastery
By the end of this stage
You can estimate what a design costs, spot the obvious waste across query engines, and go through a deploy checklist without leaving a bucket public.
Finish this track, get a certificate
Complete all 14 lessons and claim a certificate with a link you can add to LinkedIn or send to a recruiter.
Mistakes to avoid on this track
| Common mistake | What to do instead |
|---|---|
| Collecting certifications before understanding the primitives. | Certification exams test product names. Learn storage, identity, and cost first, then the exam is easy and the knowledge is real. |
| Treating a warehouse as a general purpose database. | Warehouses are built for scanning large amounts of data, not for many small lookups. Using the wrong one is slow and expensive. |
| Giving pipelines broad admin permissions because it is quicker. | It is quicker until an incident. Least privilege is the default expectation on any real team. |
Where to practise this
Where to go after this
Cloud Platforms for Data Engineers reviews & rating
Finished a few lessons? Tell us what worked and what was confusing. Reviews are read by hand and help decide what gets rewritten next.