Every technical guide we publish goes through a verification process readers never see. For our recent tutorial on safely updating n8n across self-hosted deployments, that meant more than a final proofread — it meant a DevOps engineer stress-testing every step against real infrastructure.
We sat down with Lazuardi N. Putra, Senior DevOps Engineer at Autobahn Security, to talk about his path into DevOps, how AI has changed his day-to-day work, and what it actually took to catch a handful of glitches in a runbook that had already been through several rounds of careful drafting.
What drew you to DevOps?
I actually started out wanting to be a programmer. My first IT job was as a Linux programmer, writing Bash and Python scripts to support infrastructure. Then my manager left, and I ended up taking over the server side. It gave me a different perspective: developers are the rockstars of the show, but DevOps and sysadmins build the stage. If the stage is shaky, the performance suffers no matter how good the rockstar is.
What do you like about vulnerability management?
It’s a bit nostalgic, honestly. I chose Linux early on because I wanted a different path in security — not that Windows is bad, but we all know how many Windows machines end up infected. I’ve always liked “babysitting” servers from a security angle, like building a wall brick by brick. Vulnerability management — and Autobahn Security specifically — is that same instinct applied at scale. Instead of watching one wall, you’re watching an entire fleet, making sure no brick anywhere is loose.
What’s actually changed about your work in the AI era?
Honestly, less than people think — mostly speed. What hasn’t changed is that you still act as the supervisor. You monitor the agent; you don’t hand it the keys and walk away. Before, you’d delegate to a team member and ask questions. Now you have an AI agent instead — emotionless, fast, tireless. As long as you know the goal and understand the tools, it’s genuinely useful to lean on.
Where’s the field heading?
Right now I’m sharpening my own skills — Kubernetes and CI/CD — and still architecting that same rock-solid wall. Zooming out, AIOps is what I’m watching closest: using AI to catch anomalies and predict issues before they become incidents. Platform engineering and GitOps are also gaining serious momentum. And security is baked into all of it now — DevSecOps, running checks inside the CI/CD pipeline itself rather than bolting them on afterward, is table stakes, not a nice-to-have.
Walk us through catching the glitches in the n8n runbook.
I validated it AI-assisted, but not blindly. I know Docker, Kubernetes, and my way around a Linux shell, so the process was: have the AI walk through and verify the runbook, reproduce the steps locally with AI assistance, then step in manually the moment something looked off — like a case where the AI flagged that a command’s arguments were wrong.
That’s the value of a human reviewer even after careful AI drafting: the AI can draft and self-check, but it takes someone who’s actually run the commands before to catch the moment something doesn’t smell right.
Shares to Social Media
