Case studies

Work

Each case study covers the problem, my role, the architecture and the measured outcome.

AI

AI · Amplify-Now · independent research

Cloud orchestrates, local executes: local-LLM research

Can local open-weight models safely take over mechanical AI work? I ran a structured evaluation of 19 model configurations and several agent harnesses, and applied the same pattern to an "AI factory" for coding and customer engagement.

25–50% projected token saving

AI · WA Department of Mines × DataDivers

Asking a geoscience database questions in plain English

A proof of concept for the WA Department of Mines that put a graph database and Power BI natural-language AI over public petroleum, mining and seismic data. Users could type questions like "show me all wells deeper than 500 m".

Natural language queries over geoscience data

AI · WA Health Hackathon 2023 (WADSIH × WA Department of Health)

EDDIH: Technical Achievement Prize at the WA Health Hackathon

An Emergency Department Digital Intelligence Helper for WA Health's ER-staging challenge, built on synthetic health data in 10 days. It won the Technical Achievement Prize (2nd place).

2nd Technical Achievement Prize

Leadership

Data platforms

Data platforms · INX × DataDivers

Project Bifrost: an API without an API

A metadata-driven integration on AWS S3, Airflow and Snowflake. INX's customers can import and export their own data safely, without anyone babysitting the process. Presented at the Snowflake User Group Perth.

~0.6 FTE saved on manual integration work

Data platforms · Telethon Kids Institute × DataDivers

ORIGINS: the pipeline that became a research data platform

A 2-week POC automating Telethon Kids Institute's monthly ORIGINS reporting turned into a multi-year partnership and the foundation of the institute's Data and Analytics Platform.

2 weeks to automate a 3–5 day monthly process

Data platforms · Arc Infrastructure × DataDivers

Real-time rail-sensor monitoring across Western Australia

Arc Infrastructure's in-house monitoring product was replaced with a real-time pipeline. Level-crossing sensor data flows from collectors to TimescaleDB, an Azure medallion architecture and a live Power BI status and alarm report.

Live status and alarms for every sensor and location