AI Lab · 01

Smart Rodent Control.

An IoT operations platform that helps a city’s pest-control team find rat activity early, send the right operator and prove the work was done. Designed and vibe-coded end to end by me with Claude — no developer, no Figma file.

Supervisor console: map of Antwerp with every bait station by status, and the queue of alerts that need attention

Operator field app: today’s tasks for Lotte, the operator in Antwerpen

Role

Product Builder · strategy, UX/UI, brand, design system and the build

Timeline

2019 concept · 2026 product

Built with

Claude, React, TypeScript, MapLibre, real Flemish open data

Recognition

1st Prize, Antwerp & Helsinki · Select For Cities, 2019

The problem

Antwerp doesn’t have fewer rats. It has less data.

Cities fight rats reactively: by the time someone writes it down, the colony is established. SRC closes the gap with 82 connected bait stations across seven districts, turning bait, battery and activity signals into prioritised work.

Brown-rat records in Flanders’ open data: 6,147
Of them from Antwerp, 2016–2024: 15
Antwerp’s rank of 255 municipalities: #72

How I built it

One designer, no developer.

I directed in plain language, Claude wrote the code, and I reviewed every iteration live in the browser until it met my bar. A working product shows the edge cases a static mock hides.

01Brief
02Direct
03Review live
04Decide
05Ship

I ownedProblem framing and product brief
Claude handledWriting the code
I ownedInformation architecture and flows
Claude handledWiring the map, charts and data
I ownedVisual design, brand and design system
Claude handledImplementing components to spec
I ownedEvery product and UX decision
Claude handledWriting tests for the rules I defined
I ownedQuality bar: what ships and what doesn’t
Claude handledFixing what I flagged

Two users, two products

Same data. Different jobs.

A supervisor in an SRC polo shirt works on a laptop in a city office

Supervisor console

Pest-control supervisor at the city

“Where is the risk, and who should handle it?”

Device: Desktop, office
Sees: The whole network and every team

A field operator in an SRC jacket checks the app on his phone beside an open bait station on an Antwerp street

Operator field app

Field operator in the Antwerpen district

“What do I do next, and how do I prove it’s done?”

Device: Phone, outdoors, one hand
Sees: Only their own work

Key design decisions

Built for real people.

Map legend: Offline 3, Critical 3, Attention 9, Healthy 67

Also: every number shows its source — sensors, open data or derived · status never relies on colour alone · operators see only their own work.

Outcome

Two products, one loop, live.

Two connected products covering the full loop: detect, prioritise, dispatch, act, prove. It runs on real Flemish open data. It started as a first-prize concept by a team of three in 2019; in 2026 I rebuilt it alone, from scratch.

Live · desktopSupervisor console Live · phoneOperator field app

Behind the work

Want the walkthrough?

Both products are live, so you can try them yourself. I’d also be happy to walk you through the product decisions, the build with Claude and the brand in an interview.

Let’s talk