DevOps Engineer Portfolio

I'm Hariom Singh, a DevOps engineer. This portfolio shows the systems I've built and run, from a production platform hosted single-handedly to containerised stacks and automated deployments, grouped by what each one demonstrates.

What is a DevOps engineer portfolio?

A DevOps engineer portfolio is a collection of real projects that shows how an engineer builds, ships and runs software. Unlike a resume, it links to working systems, source code and write-ups that explain the decisions behind them. Hiring teams use it to check six things: CI/CD pipelines, infrastructure as code, containers and orchestration, cloud and edge deployments, monitoring and alerting, and security built into delivery.

The sections below follow those six areas. For each one I show the project that demonstrates it, the stack, and where to see the code or the running system.

Flagship project: hosting the Saarthix platform solo

Saarthix needed a production platform with no infrastructure team and no budget for a large managed setup. I designed, deployed and run the whole thing myself. Traffic enters through Cloudflare (DNS, CDN, TLS and WAF), passes a load balancer and an API gateway that act as the single front door, and reaches containerised Spring Boot and React services. A CI/CD pipeline ships every change, and monitoring and alerting let one person keep it healthy.

The full write-up covers each layer, the delivery pipeline, observability, security and how I kept costs down: read the Saarthix DevOps case study. You can also try the interactive hosting map on the home page, which simulates traffic spikes, failovers and deploys.

CI/CD pipelines

What it shows: changes reach production automatically, safely and repeatably. This portfolio is the public example. It's plain HTML, CSS and JavaScript in the Portfolio-devops repository, and every push to main deploys it to Cloudflare in under a minute, with no manual steps.

• Config as code: security headers, cache lifetimes and redirects live in _headers and _redirects files in the repository, so every change is reviewed in Git.
• Cache safety: stylesheets and scripts are versioned on each release, so returning visitors never get new HTML with old CSS.
• Search delivery: a sitemap, structured data and IndexNow notify search engines when pages change.

For Saarthix, the pipeline lints, tests, scans, builds and deploys every change. The pipeline section on the home page walks through each stage.

Infrastructure as code

What it shows: environments are described in files, not clicked together in a console, so they can be reviewed, versioned and rebuilt. In the Saarthi-Test stack the entire environment (three services, their build contexts, ports, start-up order, health check and data volume) is declared in a single docker-compose.yml. On this site, hosting behaviour is versioned in the same way. I also work with Terraform and Ansible; you can see the full list in the skills section.

Containers & orchestration

What it shows: applications packaged once and run the same way everywhere. Saarthi-Test is a test-hosting version of SaarthiX with three containers: a React (Vite) frontend served by Nginx, a Spring Boot (Java 17) API, and MongoDB 6. The API waits until the database passes its health check before it starts, and the data lives on a named volume so it survives restarts.

backend-test:
  build: { context: ./backend, dockerfile: Dockerfile }
  depends_on:
    mongodb-test:
      condition: service_healthy
mongodb-test:
  image: mongo:6
  volumes: [ mongo-test-data:/data/db ]
  healthcheck:
    test: mongosh --eval "db.adminCommand('ping')" --quiet
    interval: 10s
    retries: 5

Excerpt from saarthix-test/docker-compose.yml.

Cloud & edge deployments

What it shows: real systems serving real traffic. Saarthix runs behind Cloudflare, which handles DNS, TLS, caching and the WAF, so most static traffic never reaches the origin. This portfolio is served from Cloudflare's edge with Brotli compression, HSTS and long cache lifetimes for fonts and images. I've also packaged a customer lifetime value model as a Streamlit app with a Procfile and setup script, so it deploys to a platform-as-a-service with a single command.

Monitoring & performance

What it shows: measuring systems instead of guessing. For Saarthix, monitoring and alerting are what make it possible for one person to run the platform. On this site, I measured performance with Lighthouse and fixed the causes rather than the symptoms: forced layouts at start-up, render-blocking third-party fonts and layout shifts. The mobile performance score on the live site went from 35 to between 75 and 86 in my lab runs, and layout shift dropped from 0.238 to under 0.04.

Security built into delivery

What it shows: secure defaults, applied automatically. This site sends HSTS, X-Content-Type-Options, Referrer-Policy, X-Frame-Options and a restrictive Permissions-Policy from version-controlled config, and redirects internal files away from public URLs. On Saarthix, the edge WAF and rate limiting stop abusive traffic before it costs anything, and the API gateway enforces authentication for every route.

More projects

Beyond DevOps, I've built web apps, machine learning models and Android apps, including a house price model that compares regressors in a scikit-learn pipeline, the TraCo COVID-19 tracker, and an Android SMS app in Kotlin. You can filter them all in the projects section on the home page.

FAQ

← Back to the interactive portfolio