Prioritising when everyone is right
Engineering wants to rebuild it, risk wants a control, security wants the dependency gone, the business wants what was promised. Scoring frameworks do not help, because you choose the weights.
I turn complex AI, data, platform and governance problems into prioritised, governed, production-ready capability inside complex organisations.
Most organisations are not short of ideas or pilots. They are short of the connective work between a promising proof of concept and something the business can actually run, fund and rely on. That work is part product, part platform, part governance, and mostly translation.
I currently own the enterprise machine learning platform at a New Zealand bank: the strategy, the roadmap, the squad, and the path a model takes from someone's experiment to a service the organisation depends on. I write here about what that path really costs, and how to make it shorter without making it reckless.
Writing
Frameworks
Prioritisation
Decision Under Constraint
How to prioritise when engineering, risk, compliance, security and the business all want different things and all of them are right.
Platform operating models
The Platform Responsibility Line
A platform that owns too little is ignored. A platform that owns too much becomes the bottleneck it was built to remove.
Governed adoption
Governance as a Path
Control functions are not the obstacle. Making every team rediscover the route through them is.
Selected work
Governance · Operating model
Designing a production gate across six control functions
Every team was rediscovering the same route through governance. Building the route once, without taking authority away from anyone.
Platform adoption · Operating model
Improving platform onboarding as a product problem
Getting a new team productive on a shared platform took weeks. Almost none of that was technical setup.
Enterprise ML · Regulated banking
Taking a machine learning capability from experiment to production service
A well-performing model that no part of the organisation was structured to run. What it took to close that gap, and what it cost.
Experience
Technical Product Owner (Service Owner), Enterprise ML Platform — Westpac New Zealand · 2024 to present
Own the strategy, roadmap and delivery of the bank's enterprise machine learning platform. Lead a cross-functional squad of five specialists and partner with data science and business teams across the organisation to move use cases from experimentation into governed production.
Data Engineer & Product Owner, Enterprise ML Platform — Westpac New Zealand · 2021 to 2024
Built and evolved the bank's data science workbench into a shared, governed platform. Delivered core AWS data and machine learning pipelines, and embedded monitoring, CI/CD and compliance capability in partnership with Security, Risk and Privacy.
Co-Founder & Product Manager — Sixth Official · 2020 to 2021
Sports technology startup. Led discovery, MVP definition and pilot delivery of a performance analytics product, working directly with coaches to validate real needs. Ran a live pilot with an amateur football club.
Data Scientist — The Clinician · 2020 to 2021
Digital health technology. Delivered analytics and reporting used by clinical teams to track patient outcomes, translating clinical requirements into practical data products.
Earlier — Department of Corrections NZ, Wellington UniVentures, Contherm Scientific · 2013 to 2020
Database development, product development and software engineering. The technical foundation the rest of this is built on.
Books and ideas
Carlota Perez
Technological Revolutions and Financial Capital
Why the value in a technology shift lands decades later, with whoever rebuilds the institutions around it.
Donella Meadows
Thinking in Systems
Short, plain, and the reason I look for the feedback loop before I look for the person who made the mistake.
Brian Christian
The Alignment Problem
A history of the distance between what we specify and what we meant, written before that became a procurement question.