Search our candidate base for "Cloud Computing" and you find 56 profiles. Search for AWS and you find 3,806. Dividing one by the other, the brand name is listed about 68 times as often as the general term.
This is the first lesson in hiring a DevOps or cloud intern: students describe themselves by tool. If your posting or your search uses a category word, you will miss most of the people who could do the work.
What the infrastructure pool looks like
The numbers are from MyInternships.in profiles on 1 October 2026. They capture skills that candidates typed in themselves; we have not examined anyone, and listing several skills places a candidate in several rows.
| Skill listed on profile | All India | Bengaluru | Pune | Hyderabad | Chennai | Delhi |
|---|---|---|---|---|---|---|
| AWS | 3,806 | 450 | 376 | 330 | 268 | 208 |
| Linux | 3,634 | 482 | 379 | 326 | 243 | 184 |
| Docker | 3,110 | 360 | 313 | 213 | 212 | 196 |
| Git | 10,194 | 1,200 | 914 | 889 | 652 | 536 |
| Cloud Computing | 56 | 8 | 6 | fewer than 5 | 6 | 6 |
AWS, Linux and Docker cluster close together: 3,806, 3,634 and 3,110. Bengaluru and Pune have the deepest pools for all three on our platform. In Pune, Linux is listed by 379 candidates and AWS by 376, practically level. In Bengaluru, Linux at 482 is ahead of AWS at 450.
Git, at 10,194, is listed roughly three times as often as any of them. That makes sense. Version control is taught in most programmes now, while infrastructure skills are still mostly self-taught.
The self-taught signal
Few colleges give students a server to manage. A fresher who lists Linux, Docker and AWS usually got there through curiosity: a home lab, a cloud provider's student tier, a hackathon where something had to be deployed at two in the morning.
That history is your best screening material. In the first call, ask "Tell me about something you deployed and what went wrong." Real experience produces a specific story with a specific error. Borrowed experience produces a list of service names.
Be wary of profiles that list a long row of cloud services with no project beneath them. A certification is a reasonable start, but on its own it proves study, not operation.
What an intern can and cannot own
Be honest with yourself before you post. A DevOps intern should not hold production credentials in the first month, should not be the only person who understands a pipeline, and should not be on call alone.
What they can do well:
- Write and tidy Dockerfiles and compose files.
- Document how environments are set up, which often exposes gaps nobody had noticed.
- Build small automation scripts for repetitive tasks.
- Improve a staging pipeline under review.
- Set up dashboards and alerts that a senior engineer then tunes.
Put these in the posting. Specific duties attract candidates who have done similar things, and they discourage the ones chasing a title.
The practical test: fix the container, write the runbook
You do not need to hand out a cloud account. Prepare a small repository with a simple web application and a Dockerfile that has three faults planted in it: a wrong base image tag, a missing dependency step, and a port that is not exposed correctly.
The instructions:
- Make the application run in a container on your own machine.
- Write down each problem you found and how you diagnosed it.
- Write a one-page runbook that a new teammate could follow to start, stop and check the service.
- List two things you would change before this went to production.
Give a weekend. The work takes a capable student about two hours, and the unhurried window respects those with exams.
Judging the submission
The diagnosis notes are worth more than the fix. "The build failed at the install step, the log said the package was not found, so I checked the requirements file" shows a method. "Fixed errors" shows nothing.
The runbook reveals whether the candidate can write for another person. Operations is largely communication: clear steps, exact commands, what normal output looks like.
The "two things I would change" answer separates levels. Common good answers include running as a non-root user, pinning versions, adding a health check and keeping secrets out of the image. You are not checking for a fixed list. You are checking that the candidate knows a working container and a production-ready container are different things.
Then spend fifteen minutes on a shared terminal. Ask them to find which process is using a port, read the last lines of a log, and check disk usage. Basic Linux fluency shows within minutes and cannot be faked.
A safe first month
Create a sandbox account or project with a spending alert, and tell the intern its limits. Cloud bills from forgotten resources are a classic first-month accident, and preventing them is your job, not the intern's.
Review every change. Explain why a rule exists when you enforce it. Interns who understand the reason behind a guardrail keep to it when nobody is watching.
For more on structuring technical internships, browse our employer resources and employer tools. To reach candidates who list these tools, post a DevOps internship on MyInternships.in and name AWS, Docker and Linux in the title, not only "cloud".


