A governed data platform for enterprise reporting
Led the architecture and delivery of a governed lakehouse and warehouse, with data quality checks and a named owner for each domain.
Data & AI leadership · London
AI that holds up in regulated work.
I lead data and AI programmes in regulated industries, from governed data platforms to generative AI in production, with delivery for FTSE 100 organisations.
Regulated by background
I started in medical-device quality, where a wrong number is a physical failure, not a slide.
Enterprise delivery
Governed data platforms and AI tools taken into production for FTSE 100 organisations in regulated sectors.
Hands-on
Close enough to the code to catch a bad architecture decision early. Senior enough to own the outcome with the board.
What I believe about AI now.
-
Pilot
A model that demos well.
-
Validated
Shown to behave the same way every time.
-
Owned
A named owner and a measure they already report.
-
In production
Monitored, audited and used.
Pilots are cheap. Validation is the work.
Anyone can demo a model. The value sits in showing it behaves the same way every time, inside a process an auditor or a regulator will inspect.
AI changes the shape of the team, not the need for judgement.
Agents take on more of the routine work. What remains is senior: deciding what to trust, and owning the result.
If nobody owns the number, the model doesn't matter.
Every data and AI investment needs a business owner and a measure they already report on, agreed before the work starts.
From governed data to AI in production.
Examples from my current role, described at the level of the problem and my part in it.
One governed view of security risk
Designed and led delivery of a security asset intelligence platform that brings the security estate into one governed model for leadership reporting.
Safe server decommissioning during a cloud migration
Architected an engine that maps service dependencies and checks each decommissioning candidate before anything is switched off.
Generative AI for drafting business cases
Built a retrieval-augmented generation tool that drafts business cases against the organisation's own criteria. People still own the final case.
An engineer's standard for what counts as a result.
Control engineer by training, medical-device quality by experience. That taught me one standard I still apply to data: if a measurement doesn't change a decision, it isn't worth producing. Since 2021 I've led enterprise data and AI work for FTSE 100 and regulated-sector organisations.
MSc Artificial IntelligenceDistinction
Azure Solutions Architect ExpertPlus data engineering, AI engineering and data science certifications
Executive Strategy ProgrammeSaïd Business School, University of Oxford
Peer-reviewed researchSix papers through IEEE, IET and MDPI Energies
Always glad to compare notes.
If you work on data and AI in regulated industries, I'd be glad to hear how you're approaching it.