After years of sharing thoughts on Twitter in fragments, I’m starting a proper home for longer-form technical writing. Welcome to zedzed.dev.

What this is

This site is a content brand, not a personal diary. You’ll find:

• Cloud-native infrastructure — Kubernetes patterns, platform engineering, eBPF, GitOps workflows, and the operational realities nobody puts in the docs.
• AI tooling — How AI agents actually work in production, what LLM-powered developer tools get right (and wrong), and practical patterns for integrating AI into engineering workflows.
• SRE insights — Reliability engineering from the trenches: incident analysis, observability patterns, toil reduction, and how to build systems that don’t wake you up at 3 AM.

What this is not

This is not a “here’s how to do kubectl get pods” beginner tutorial site. Plenty of those exist. My target reader is an engineer who already knows the basics and wants to go deeper — someone who has run Kubernetes in production and still has questions.

Why now

Three reasons:

1. AI changed the tooling landscape faster than anyone could blog about it. I want to document what actually works in production environments, not what the press releases say.
2. The SRE discipline is maturing. The conversations have gotten more interesting. What does “reliability” mean when your LLM inference infrastructure is serving 10M requests a day?
3. Writing forces clarity. If I can’t explain why a system behaves a certain way, I don’t understand it well enough.

Frequency

I’ll publish when I have something worth saying. No arbitrary cadence, no filler content. Subscribe via RSS or email if you want updates without checking back.

The first real technical post is coming soon. It’s about something I’ve been thinking about for months: how platform teams should think about AI agent infrastructure differently from traditional microservices.

See you then.

— STRRL