TL;DR: A strong DevOps portfolio has six projects: (1) a manual deployment on a VM, (2) the same stack in Terraform, (3) a CI/CD pipeline with a test gate, (4) Kubernetes with a deliberate failure you diagnose, (5) monitoring plus an incident write-up, and (6) a cost and security clean-up. Each README should explain what broke and how you fixed it. Guided versions of these projects are on Bartman.

Certifications get you past the filter. Projects get you the offer. When I review DevOps candidates, I don't want ten "hello world" pipelines. I want a few projects that show you've hit the problems real teams hit.

What should a DevOps portfolio include?

1. How do you deploy a full-stack app on a single VM?

Take a MERN (or any) app and put it on an EC2 instance with Nginx as a reverse proxy, a process manager, HTTPS and a domain.

Proves: you understand what the automation later hides. Ports, security groups, DNS, certificates, logs.

2. Why rebuild it with infrastructure as code?

Same app, now provisioned by Terraform or OpenTofu: network, instance, security groups, state in a remote backend.

Proves: you can make infrastructure repeatable and reviewable.

3. What makes a CI/CD pipeline portfolio-worthy?

Jenkins or GitHub Actions: build, test, scan the image, push to a registry, deploy. Make a failing test block the release and show it in the README.

Proves: you think about what should stop a deploy, not just what triggers one.

4. Why should your Kubernetes project include a failure?

Deploy to a cluster with health checks, resource limits and a rolling update. Then break it on purpose (bad image tag, OOM, failed readiness probe) and document how you diagnosed it.

Proves: troubleshooting. This is the part interviewers dig into.

5. What does observability look like in a portfolio?

Add metrics, logs and one alert that matters. Trigger it, then write a short post-incident review: timeline, cause, fix, prevention.

Proves: you can operate a system, not only ship it.

6. What is a cost and security pass?

Find the waste (idle resources, oversized instances, unattached volumes) and the risks (public buckets, wide IAM policies). Fix them and record the before and after.

Proves: you think like the person who pays the bill and answers the audit.

How should each project's README be structured?

1. The problem, in one paragraph.
2. An architecture diagram.
3. How to run it.
4. What broke and how you fixed it.

That last section is what makes a portfolio memorable.

Key takeaways

• Six focused projects beat twenty tutorial clones.
• Each project should prove one distinct skill: deploying, automating, gating, troubleshooting, operating, optimising.
• Break things on purpose and document the fix; that's what interviewers ask about.
• A clear README matters as much as the code.

FAQ

How many projects does a DevOps fresher need? Four to six well-documented projects that cover deployment, infrastructure as code, CI/CD, containers and monitoring are enough for most junior roles.

Do I need to pay for AWS to build these? Most of it fits in the AWS free tier if you tear resources down after each session. Set a billing alert before you start.

Are certifications or projects more important for DevOps jobs? Certifications help you pass résumé filters; projects are what you discuss in the interview. Have both, but lead with projects.

Where can I practise DevOps with guidance?Bartman has step-by-step versions of these projects with an AI mentor and a review of your submission, plus a DevOps and Cloud Engineer simulation that runs incidents and change requests against your own AWS account.

Where to go from here

I run Bartman, where most of these exist as guided projects, from deploying a MERN app on EC2 to Jenkins pipelines for Kubernetes and FinOps on AWS. But you can build all six on your own. The important part is building them.

Written with help from AI tools.