About
## What I work on
Enterprise AI implementation
How an organisation gets from scattered tools and unclear data to AI its staff can trust: finding where work and data actually flow, who owns them, and what has to be true before an assistant can answer questions safely.
- Enterprise AI ImplementationModeled
An end-to-end implementation method worked through on a fictional organisation: discovery, data ownership, pipelines, data quality, and a secure question-answering layer. Modeled, not a client deployment.
- AI Workflow Discovery AutomationUnavailable
An n8n and ChatGPT workflow that asks staff how they use AI, which tasks recur each week, and how long they take, then points them toward possible solutions. Built and previously live; offline until its hosting is restored.
AI-native product engineering
Building real software with AI in the loop while keeping the engineering disciplines that make it dependable: planning, security, tests, documentation, and human verification of everything the AI produces.
- WhakapapaIn development
A fast, collaborative way for Māori users to build, trace, and share family trees, on a visual canvas rather than a form-heavy database. In development.
- Kupu MāoriLive
A live Māori dictionary application, shipped and iterated with an AI-native workflow. Dictionary data is currently failing to load; the fix is in progress.
## How I work with AI
- AI drafts, researches, and writes code; I own the decisions, review the output, and verify it before it ships.
- Workspaces and permissions are set up deliberately, with secrets kept out of reach of the tools.
- Each case study shows the process — planning, tests, failures, and fixes — not just the finished screen.
## Background
Full-stack development training, AWS Certified Cloud Practitioner, and an internship in Datacom’s Emerging Technologies team. Before that: marketing systems, technical facilitation, and consulting. That work was mostly about helping people and organisations adopt technology, which is the same problem AI now raises.