How DevOps and AI Opened Doors Remote Work Never Would Have

How DevOps and AI Opened Doors Remote Work Never Would Have

There was no plan for any of this.

I started out as a full-stack developer, building the kind of applications everyone tells you to build when you're learning: a Next.js frontend, a FastAPI backend, MongoDB tucked in behind it, all wrapped in Docker because someone on a forum said that's what "real" developers do. I didn't know yet that I was standing at the edge of something bigger. I just knew I wanted to understand what happened after the code where it actually lived, how it stayed alive, what broke it.

That question is what pulled me into DevOps. And DevOps is what, eventually, pulled me into AI.

The pivot nobody warns you about

Self-taught paths don't come with a map. I learned Kubernetes the way I learned most things — by breaking it first. Kube-prometheus-stack, Grafana dashboards I didn't fully understand yet, ArgoCD running on a Kind cluster on my own machine, late at night, with no mentor checking my work and no classroom validating whether I was doing it "right." I built a GitOps promotion pipeline — Kargo and ArgoCD moving a small todo app through dev, staging, and production on a k3s cluster not because a job required it, but because I needed to know, for myself, that I could.

Somewhere in that process, agentic AI stopped being a buzzword and started looking like the next real question. If I could teach infrastructure to promote itself safely, could I teach a system to notice when something was wrong, reason about it, and know when to stop and ask a human? That question became my AI Logging Agent a multi-agent system built with the OpenAI Agents SDK, Temporal for durable execution, and human-in-the-loop checkpoints baked into the remediation logic, because I never wanted to build something that acted with more confidence than it deserved.

None of this was glamorous while it was happening. It was a lot of 11 p.m. debugging sessions, a lot of "why is this pod still pending," a lot of quiet frustration with no one in the room to share it with.

## What the skills actually bought me

Here's the part that surprised me: none of this was built for opportunity. It was built out of stubbornness. The opportunity came after, almost as a side effect of taking the work seriously enough to make it visible.

A Kubernetes competency center reached out to me directly a real client inquiry, unprompted, because they'd seen the infrastructure work I'd been sharing. I went through a full DevOps assessment process with Crossover a cognitive aptitude test, a complex problem-solving stage, and a timed root-cause-analysis challenge on an OpenSearch CPU saturation incident, submitted through a custom tool inside Cursor, no safety net, just me and a clock. I didn't get there because of a degree or a name on a resume. I got there because the skills were demonstrable, and demonstrable skills travel — across time zones, across borders, past the parts of a hiring process that usually gatekeep people who look like me on paper.

That's the real shift DevOps and AI gave me. Not a job title. Portability. The kind of work where what matters is whether the pipeline runs, whether the agent makes the right call under pressure, whether the RCA holds up — not where I'm sitting when I do it.

The part I don't want to skip

I don't want to sell this as frictionless, because it wasn't.

Remote opportunity doesn't erase the isolation of being self-taught in a field that still, quietly, prefers credentials. It doesn't erase the extra hours I spent proving I understood why something worked, not just that it did, because no one was going to take that on faith from someone without the usual pedigree. Being based in Karachi and building for clients and communities on the other side of the world means the opportunity is real, but so is the timezone fatigue, the sense of always translating your work into a language — professional, technical, cultural that wasn't built with you in the room.

I think that's the honest version of "opportunity." It's not a door swinging open. It's a door that opens if you keep showing up with something real to put in front of it.

Why I still believe it

I believe DevOps and AI are some of the most genuinely portable skill sets available right now not because they're trendy, but because the work speaks for itself in a way few other fields allow. A pipeline either promotes safely or it doesn't. An agent either makes the right remediation decision or it escalates appropriately. There's less room for someone to discount you when the artifact is sitting right there, working.

That's what remote work through this field gave me that nothing else had: a way to be evaluated on the thing I actually built, not the version of me that fits neatly into a room I was never in. I didn't start this to prove a point about women in tech, or about self-taught engineers, or about working from Karachi instead of San Francisco. I started it because I wanted to understand how things stayed alive after they were built.

It just turned out that understanding was worth more, and traveled further, than I expected.