Case studies
Systems we’ve built, end to end.
Anyone can show a screenshot. What’s harder to fake is a system still running years after launch. Grouped below by what we built, not in arbitrary order.
01
Platform & data engineering
Platform & data engineering
Full-stack systems for organisations running distributed operations.
One system of record across 1,100+ dealers.
A national dealer network ran on locally-held records: each dealership held its own view of inventory, customers, service history and warranty, in whatever shape local practice had settled on. Nobody at the centre could answer a question about the network without asking 1,100 people.
Partner enablement across languages, in production for years.
A video systems manufacturer sells through a partner channel spread across markets and languages. Product knowledge, configuration guidance and enablement material lived in documents that went stale the moment a product revision shipped.
Retail execution intelligence across 6,000 stores.
A consumer goods business ran in-store audits across six thousand outlets through field teams. The audits generated more than twenty million records, and the reporting on top of them arrived slowly enough that a compliance problem was historical by the time anyone saw it.
Point-of-purchase compliance, verified by photograph.
A retail audit firm verified point-of-purchase display compliance on behalf of brands. Verification depended on field photographs, which arrived as unstructured evidence attached to inconsistent report formats.
02
Legacy modernisation & quality engineering
Legacy modernisation & quality engineering
Bringing hybrid legacy-and-modern stacks under real test automation and delivery discipline.
Framework designed · in commercial discussion, not yet delivered
We’ve designed a testing and modernisation framework for hybrid AS/400 (RPG/CL) and .NET environments — RPGUnit-based orchestration, tn5250j screen automation, Playwright for the web layer, unified reporting through Allure. We’re currently in commercial discussion with a collections/AR management firm to bring it into their environment. No case study yet — we’ll publish one once there’s real delivery behind it, not before.
03
Applied AI
Applied AI
Matching, scoring and retrieval systems that reach production and stay there.
One marketing intelligence platform, 750+ businesses, then handed over.
A marketing services firm serving hundreds of home-service businesses had performance data spread across more than twenty advertising, call-tracking, CRM and web analytics sources. Every client report was assembled by hand, which capped how many clients the business could serve without adding people at the same rate.
A candidate matching and attrition-risk engine, later productised by the client.
A large professional services organisation was matching candidates to roles through keyword search over résumés, which surfaced the wrong people and missed the right ones. Separately, they had no forward-looking read on which staff were most likely to leave.
A skill taxonomy as a graph, not a keyword list.
A recruitment marketplace needed to match people to roles where the useful signal was adjacency — this person hasn't done that exact job, but everything they have done sits one hop away from it. Relational schemas make that query expensive and the results shallow.
Next
Forge the model. Build the system. Own it.
Ninety minutes with the people who own the problem, and a written point of view you keep either way.
Book a 15-min architecture call