The unusual part of this career is the direction: up the whole stack, from analog systems to cloud products, one rung at a time. This is why the product judgment has the smell of the datacenter on it — told chronologically, with the history that rhymes. Only dated facts carry dates; the early years are honestly untimed.
A home computer, a DOS prompt, analog phone lines — and a Linux CD that arrived by international post.
It started with Windows 95 and Norton Commander: two blue panels, a keyboard, and a file system to explore. There was no internet to fall back on, so the operating system itself was the playground — install, break, reinstall, and slowly learn what a computer actually is.
Every abstraction I would later depend on professionally was, at some point in those years, something I had to physically reinstall at 2 a.m.
Before digital networks, there were analog systems to understand — wiring, signals, the physical layer that everyone forgets once it works. Hands-on time with analog equipment taught me that every 'virtual' thing eventually touches something real.
Twenty years later I still trace problems down the stack to the physical layer first. That instinct was built here.
I filled in a web form on a borrowed connection and waited. Weeks later, an envelope from the Netherlands arrived with a pressed Ubuntu Linux CD inside — an operating system shipped across a border, free, by post. I installed it, and the command line became home.
That disc was my first direct contact with the open-source world: strangers on another continent, mailing a stranger the tools to learn. Everything I've built since runs on what that envelope started.
VoIP administration at Tel4Tel: the moment telephony stopped being wires and became software.
Tel4Tel
At Tel4Tel I administered VoIP systems: call flows, SIP trunks, codecs, gateways. Telephony — the century-old empire of copper and switches — was being rewritten as applications, and I was on the team doing the rewriting for real customers.
This is where the two halves of my future met: the analog discipline of the phone network and the programmable logic of software. Voice stopped being physics and became configuration.
Technical support, network engineering, then VoIP expertise at FCP — learning the whole stack from ticket queue to trunk.
FCP
Joining FCP in technical support meant owning the moment when things break for real users. Every ticket was a lesson in how systems actually fail — and how people experience that failure.
Support taught me the most durable engineering skill I know: reproduce, isolate, verify. It's also where I learned that reliability is a form of respect.
FCP
A year later I moved into network engineering: routing, switching, firewalls, the paths packets take. From inside the network, 'the cloud' stopped being a marketing word and became routers, tables and failover plans.
Networks are the nearest thing infrastructure has to a nervous system. Once you've traced a packet across one, every abstraction above feels earned.
FCP
At FCP I grew into VoIP expertise and eventually management of the VoIP function — NGN architecture, Cisco VoIP, PBX systems, and the people who kept them alive. Designing call platforms turned into designing how a team operates.
The hardest scale-up was never the call volume; it was the coordination. Managing engineers taught me that most 'technical' problems are actually interface problems between humans.
Cloud engineering and NOC leadership at a global-scale cloud provider: from physical hosts to fleets, from fixing to coordinating.
Global-scale cloud provider
Moving to a global-scale cloud provider in 2019 meant living inside the abstraction ladder: physical hosts, virtualization, orchestration, automation. Systems I had administered one by one now existed as fleets with dashboards.
The ladder — physical, virtual, cloud, cloud native — isn't marketing. Each rung changes what 'failure' means and who can fix it.
Global-scale cloud provider
Leading a Network Operations Center meant owning incidents end to end: the escalation paths, the runbooks, the handovers, and the humans at 3 a.m. In a year when the world's traffic moved indoors, the NOC was where the internet stayed up.
Incident response is applied epistemology: what do we know, how do we know it, and who needs to know it next? Everything else is keyboard work.
Technical product management for services used by hundreds of thousands of people — turning reliability into something you can design.
Global-scale cloud provider
Since 2022 I've worked as a technical product manager on cloud services serving hundreds of thousands of users — deciding what gets built, for whom, and why, with the infrastructure discipline of the previous decade behind every choice.
A product manager's job at infrastructure scale is translation: users' needs into architecture, architecture into roadmaps, roadmaps into promises the ops floor can keep.
A pressed CD in an envelope from the Netherlands, weeks of waiting, an install that had to work the first time — because reinstalling meant another postal round trip.
terraform apply. Whole environments described in code, provisioned in minutes, torn down and rebuilt without fear. The operating system is the smallest part.
Same curiosity, different ladder. That span — from a postal CD to infrastructure-as-code — is the whole Journey in one artifact.
The datacenter era is easiest to explain as a ladder. Each rung hides the one below it — and changes who can fix what.
Physical
Bare metal. You touch what breaks.
Virtualization
One host, many machines.
Cloud
Location stops mattering.
Cloud native
Servers stop mattering.
Intelligent infrastructure
Configuration starts writing itself.