On this page
Most data work now runs on a cloud platform. Companies want people who understand storage, databases and scheduled jobs on one of the big three: Amazon Web Services, Microsoft Azure or Google Cloud. For a fresher, the goal is to understand a small set of services well and to show one working pipeline.
Files land in a bucket or container.
Code cleans and checks the files.
Clean data is loaded into tables.
The job runs every day on its own.
People read the result.
Which cloud should you pick
| Cloud | Often seen in | Storage | Warehouse |
|---|
| --- | --- | --- | --- |
|---|
| AWS | Startups, product firms, large companies | S3 | Redshift |
|---|
| Azure | Large companies, banks, firms using Microsoft tools | Blob Storage, Data Lake | Synapse |
|---|
| Google Cloud | Analytics-heavy and some product firms | Cloud Storage | BigQuery |
|---|
Look at the job posts near you and pick the cloud that appears most. If you cannot tell, AWS is a common choice. Service names differ, but the ideas of storage, compute, database and scheduling are the same everywhere. The cloud platforms track compares them.
What to learn first
- Storage: buckets or containers, folders, and permissions.
- Identity and permissions: who can access what, and why least access is safer.
- A database or warehouse: load and query tables.
- Running code: a small managed service or a virtual machine.
- Scheduling: run a job daily. See what is Apache Airflow.
- Cost basics: why storing data is cheap and scanning or running machines is not.
Do not try to learn all the services. A few, understood well, is the right size for a fresher.
Practise safely and cheaply
- Use the free tier, and read what it includes and for how long.
- Set a budget alert on day one, so you are told if spending starts.
- Turn off or delete resources when you finish. Running machines and databases can keep charging.
- Never put access keys in code or in a public repository.
- Use a separate account for learning, not one tied to a company.
A project to show
Build the small pipeline in the figure. Upload a CSV to storage. Write a script that reads it, cleans it and loads it into a table. Schedule it to run daily. Draw the architecture on one page, write down the permissions you used, and add the cost you saw. A README with the diagram and the steps is enough. The air quality pipeline project has a similar shape.
What interviews ask
- What is the difference between a data warehouse and a data lake? See warehouse vs lake vs lakehouse.
- How would you store raw data, and why keep it unchanged?
- How do you give a script access to storage safely?
- How would you make a daily load safe to rerun?
- How do you keep costs under control?
Certificates can help get past a screen, but interviewers ask about what you built. Prepare to explain your pipeline from end to end.
Where to start today
Choose one cloud, create a free account with a budget alert, and upload one small file to storage.