The craft behind the tracks
Skills
The Tracks page is about problems worth building for. This page is about the actual capability that makes those tracks credible instead of aspirational: what I've built, with what tools, and where you can go check it for yourself. Nothing below is a claim without a link attached.
AI agent engineering
Most people talk about AI adoption. I sit down and build with the tools directly: directing AI coding agents through a full build, from content architecture to a working, deployed site, rather than handing that work to a developer and reviewing the result.
LLM tooling, AI coding agents, structured content pipelines, bilingual content systems
You're reading the proof right now. This entire site, its bilingual content structure, and the interactive 3D city at /world were built through agent-directed engineering, not a template.
Walk the live space →Predictive maintenance & MJOP modeling
Multi-year maintenance planning (MJOP) is where I started in real estate operations, and it's the clearest case for AI in the industry: condition data that already exists, rarely used to predict anything. I've automated this shift for housing corporations and, earlier, built one from nothing for a Dutch municipality.
MJOP methodology, condition assessment, predictive modeling
Proven at Bryder, moving Dutch housing corporations off manual spreadsheets and onto a digital-twin-backed platform.
Read the case study →Portfolio data & reporting
Getting from "the data exists somewhere" to "the dashboard people actually open" is mostly unglamorous plumbing: auditing sources, fixing ownership, and only then building the visualization layer on top.
Power BI, data structuring, dashboard design, reporting automation
Proven at Ons Doel, cleaning up the property data a housing corporation's reporting depended on, and running daily in an international portfolio I help operate today.
Read the case study →Digital twins & simulation
A digital twin is only worth building once a smaller simulation has answered the actual question. I've built both ends of that spectrum: lightweight models that settle a decision in a week, and full platforms that need to carry years of portfolio data.
BIM, digital twin platforms, Three.js / React Three Fiber
Rolled out at Bryder as the backbone of the automated MJOP platform. The 3D city at /world runs on the same spatial-interface thinking.
Read the case study →Smart buildings & ESG
Sensor data and sustainability reporting are usually treated as separate problems handled by separate teams. Certified Smart Building Assessor training is what let me connect them: the same measurement discipline feeds both a facilities decision and a compliance report.
Smart building systems, ESG reporting, sustainability measurement
Positioned at Smartvatten, tying smart building tech to a measurable sustainability story instead of a standalone gadget.
Read the case study →Real estate operations, ground-level
Everything above only works because it's grounded in operations I've actually done: inspections, tender processes, tenant conversations, procurement rules that don't bend for a nice pitch deck. That's the part most AI-for-real-estate pitches are missing.
Property management, procurement, public-sector portfolios
The full career path, from Albert Heijn to senior transformation consulting, is on the About page, alongside the full credentials list.
Read the full path on About →
Where this goes next
None of this is a menu you order from. It's the base a track gets built on when your problem is real. If one of these areas overlaps with something you're stuck on, that's usually a good sign, not a coincidence.
Book 20 minutes, no pitch deck