I design systems that quietly scale — and I'm now teaching them to think.
I'm a platform engineer with about ten years in production, now spending most of my time on systems that run on AI. Based in the Netherlands. For most of that decade I've helped grow Shiftbase into one of Europe's leading workforce-management platforms, today used by 8,000+ businesses to schedule 250,000+ employees a day.
I was the first engineer, so over the years I've worn most of the hats: backend, architecture, team lead, platform, and for a stretch the ISO security officer role too. I designed the Google Cloud foundation the product runs on and moved the team onto Kubernetes, GitOps and infrastructure-as-code. Keeping all of that secure and compliant while it scaled across European markets came with the territory.
These days that same instinct points at AI. The tools are new; the discipline isn't. I build agentic workflows, do genuine context and skill engineering, and run Spec-Driven Development on real production work rather than throwaway demos.
What turns a slick demo into something you'd hand to real users is mostly the unglamorous part: clear specs, sane guardrails, and feedback loops that actually close.
Visionary, but grounded. I'll chase the big idea, then ship the boring version that survives Monday morning.
|
Internal tooling and pipelines decide how fast everyone else can ship. I give them the same care as anything a customer sees, because they're what the whole team builds on. |
You shouldn't have to think about it. Things deploy, scale, and recover quietly, and the on-call phone stays quiet too. |
|
Ten years in one codebase taught me the wins that last are the ones a whole team can understand and own long after I've moved on. |
Agents are powerful, but they still need specs, guardrails, and someone who's shipped real systems before. The engineering doesn't get easier. It just gets more interesting. |
claude-skills — my main playground right now.
A growing toolkit of reusable skills for Claude Code: the agent workflows, prompts and guardrails I use daily, packaged so they're repeatable instead of one-off. It's where my platform habits (composability, conventions, "make the right thing the easy thing") meet AI tooling.
If you only click one link on this page, make it this one.
Plenty more lives in private IaC & platform repos — happy to talk through it.
Always happy to talk shop — platform & cloud architecture, GitOps/IaC, scaling engineering teams, or building AI-native workflows that hold up in production. If any of that is on your mind, my inbox is open.
