Skip main navigation
/user/kayd @ devops :~$ cat about.md

About Karandeep Singh About Karandeep Singh

QR Code linking to: About Karandeep Singh

Summary

Karandeep Singh
Karandeep Singh — DevOps Engineer, Calgary, AB

I’m a DevOps engineer based in Calgary, Alberta. I’ve spent the past seven-plus years designing, building, and occasionally setting fire to production infrastructure — the kind that keeps companies running at 2 AM when the alerts fire.

What I work on

Cloud architecture, infrastructure automation, and CI/CD. In practice that means:

  • AWS — EC2, S3, Lambda, EKS, CodeBuild/CodeDeploy, CloudFormation, Route 53. Migrated a 50,000-page documentation site from a $450/month WordPress stack to S3 + CloudFront for $18/month, and moved a 5M-event/day analytics pipeline from a $8,247/month always-on architecture to a serverless one running at $847.
  • Kubernetes — EKS clusters from a few nodes to hundreds, with Helm, Karpenter, Prometheus, and the usual on-call rotation that comes with it.
  • Jenkins — pipelines, JCasC, agent fleets on EC2 and Kubernetes, and the version migrations nobody volunteers for.
  • Linux & shell tools — bash, sed, awk, grep, the GNU coreutils. The kind of work where you process 2TB of logs a month and learn very fast which patterns actually scale.
  • Go — small operational tools that replaced bash scripts when they crossed the “needs tests and arguments” threshold. A task runner, a session manager, a config tool.

Why I write here

Most online tutorials skip the failure modes. They show you the happy path, hand you a working snippet, and leave you stranded the first time it breaks at 2:13 AM in production.

Karandeep Singh
Karandeep Singh — DevOps Engineer, Calgary, AB

I’m a DevOps engineer based in Calgary, Alberta. I’ve spent the past seven-plus years designing, building, and occasionally setting fire to production infrastructure — the kind that keeps companies running at 2 AM when the alerts fire.

What I work on

Cloud architecture, infrastructure automation, and CI/CD. In practice that means:

  • AWS — EC2, S3, Lambda, EKS, CodeBuild/CodeDeploy, CloudFormation, Route 53. Migrated a 50,000-page documentation site from a $450/month WordPress stack to S3 + CloudFront for $18/month, and moved a 5M-event/day analytics pipeline from a $8,247/month always-on architecture to a serverless one running at $847.
  • Kubernetes — EKS clusters from a few nodes to hundreds, with Helm, Karpenter, Prometheus, and the usual on-call rotation that comes with it.
  • Jenkins — pipelines, JCasC, agent fleets on EC2 and Kubernetes, and the version migrations nobody volunteers for.
  • Linux & shell tools — bash, sed, awk, grep, the GNU coreutils. The kind of work where you process 2TB of logs a month and learn very fast which patterns actually scale.
  • Go — small operational tools that replaced bash scripts when they crossed the “needs tests and arguments” threshold. A task runner, a session manager, a config tool.

Why I write here

Most online tutorials skip the failure modes. They show you the happy path, hand you a working snippet, and leave you stranded the first time it breaks at 2:13 AM in production.

Articles here follow a different format:

  1. Start with a real problem from production.
  2. Try the simple solution.
  3. Watch it fail in a specific, documented way.
  4. Show the fix and why it was wrong before.
  5. End with the version that actually goes in the runbook.

If that means an article on sed is mostly about the time the GNU vs. BSD difference broke a deploy on a developer’s MacBook — that’s the version I’d want to read.

Where I’ve been

Calgary’s DevOps community is smaller than Silicon Valley’s, but the work is real and the stakes are real. I’ve worked across financial services, SaaS, and infrastructure operations roles — environments where the system going down costs measurable money per minute, and where “we’ll fix it next sprint” is rarely an acceptable answer.

Connect

Reach out if you want to talk DevOps, hire help with cloud architecture or CI/CD, or just argue about whether awk is the right tool for a given task (it usually is).