Agentic AI · Enterprise Engineering · Sydney

I help enterprises
turn AI agents into
production systems.

I assemble working AI agent prototypes on compressed timelines and partner with architects to fit them into existing identity, security, and integration landscapes.

Currently shipping agentic AI in production at King Living. Background spans Cloudflare-native rapid prototypes, hybrid identity, and enterprise security architecture.

Email me → GitHub ↗ LinkedIn ↗

Currently

Cloud Engineer
at King Living

Recent PoC

1 day build
demoed next

Working with

Google ADK · Azure
Cloudflare · Gemini

01 / Selected Work

Things I've shipped
or prototyped recently.

01 2025 — present
Production system

IT Portal & Agentic Onboarding

An agentic AI system for IT operations, in production at King Living. Google ADK agents triage tickets, orchestrate onboarding/offboarding, and execute identity-aware actions across Microsoft Graph, Jira, and hybrid AD — with human-in-the-loop approval at the points that matter.

The interesting part isn't the agent. It's how the agent fits into existing enterprise tooling rather than replacing it.

Architecture Agent reasons in the cloud, proposes actions, and waits for human approval before executing on hybrid infrastructure.

Google ADKAzure FunctionsMicrosoft GraphJiraHybrid AD

02 2026
Client prototype

★ Featured · 1-day build

Recruitment Scout

A working AI recruitment prototype, built end-to-end in a single day for a New Zealand recruitment firm and demoed the next.

Cloudflare-native ingestion, LLM-enriched listings, vector search with structured reranking, and ElevenLabs voice-based candidate screening — all of it operational in 24 hours.

The point of a one-day prototype isn't the prototype. It's letting the customer see their future-state stack working, and shaping the conversation that follows.

Cloudflare Workers AID1VectorizeElevenLabs

Read the case study → Try the walkthrough →

03 2026
Internal AI system

Customer Research Assistant

A grounded RAG assistant used by an internal marketing team for persona simulation, research analysis, creative brief generation, and cited answers across customer segmentation research. Built on Gemini File Search with a Streamlit interface and adopted for weekly use.

Gemini File SearchStreamlitGrounded RAG

04 Personal project
Live product

Ready Kiwi

A New Zealand hazard information product I built end-to-end: Flutter client, Cloudflare Workers backend, AWS runtime, and AI-driven earthquake detection plus notification summaries running on a realtime detector pipeline.

Applied AI in a domain where latency and reliability are non-negotiable. Different problem shape from enterprise agents — same systems thinking.

FlutterCloudflare WorkersAWSAI detection

02 / Approach

How I think about
shipping agentic AI.

— 01

Prototype first, slides second.

A working demo on day three changes the conversation more than a forty-slide deck. I start by building the smallest version of the thing that could possibly work, then iterate against real constraints.

In practice — Recruitment Scout went from cold start to client demo in 24 hours: ingestion, vector search, reranking, voice screening. The scoping conversation that followed was sharper than any spec doc would have produced.

— 02

Fit, don't replace.

Enterprise AI agents earn their place by integrating with existing identity, ticketing, and approval flows — not by demanding green-field architecture. I design for the systems that already exist.

In practice — On the IT Portal, the agent proposes actions but executes through the existing approval flow and hybrid AD perimeter. Change management mattered more than novelty.

— 03

Honest about what won't work.

The fastest way to lose trust is to oversell. I'd rather scope a system down and ship it cleanly than scope it up and miss. Every architecture has tradeoffs, and the useful thing is to name them early.

In practice — This research assistant started as a far broader idea. We narrowed it to grounded RAG over segmentation research because that was the version we could ship and trust. It's still in weekly use.

03 / Background

Where I've been.

2025 — Present

Cloud Engineer King Living

Sydney, AU

2023 — 2024

Release Manager — Environments & Delivery Department of Internal Affairs

Auckland, NZ

2022 — 2023

IT Infrastructure Engineer Wine-Searcher

Auckland, NZ

2021 — 2022

Network Operator Auckland Transport

Auckland, NZ

2014 — 2021

Senior Logistics Specialist / Team Leader Fullers360

Auckland, NZ

Master of Entrepreneurship, University of Otago · BA, Auckland University of Technology · HashiCorp Terraform Associate · Microsoft Azure Fundamentals · AZ-104 in progress

04 / Contact