The hard calls behind what we build. And what we don't.
Each came down to a hard call before any build — what the data really needed, and what it would take to make it trustworthy. Most of these we built and still run; one we were trusted to validate, not build, when the stakes were too high to guess. Every one is still relied on, most for a decade, because the judgment came before the code.
A seabed-habitat platform marine managers still rely on
Seamap Australia
Australia's seabed data was scattered across institutions, each classifying it its own way — one team's 'rocky reef' was another's 'hard substratum.' IMAS synthesised it into a single national classification; the hard part was making that trustworthy in public hands — we built the portal, funded by Parks Australia, that put it to work. A decade on, marine-park zoning, fisheries limits and state-of-environment reporting are decided on it, not guesswork — and it won two 2018 iAwards.
10 years live Public platform
A decade running the university's estate
University of Tasmania
Running the estate meant manual processes and asset data scattered across disconnected systems — no single view to plan or spend from. We consolidated it into one platform, working with the tools they already ran rather than replacing them, and put live data in front of staff and contractors on iOS and Android. Connecting the data is what unlocked the value. It's still the system they depend on.
10+ years live Enterprise
AI seagrass mapping a ranger can run
Blue carbon science
A national science agency had AI models for classifying seagrass from underwater photos — but they stayed locked in expert pipelines, no use out in the field. We built the last mile: a validate-as-you-go importer, a regional data hub, and mapping that turns off-the-shelf camera photos into verified habitat maps. Now researchers and rangers across the Indo-Pacific monitor blue carbon themselves, and target restoration where it counts.
In production AI-assisted
The data layer research runs on
IMAS
Marine science runs on data and logic that has to stay correct, described and findable for years, or it stalls. For a decade we've grown and maintained the research databases at IMAS's core, published their data into the national and global networks, and built the portals and maps that put it in front of the people who use it. So the research keeps moving, the data outlives its project, and the advice built on it holds up.
Ongoing Research infrastructure
Brought in to make a regulator's data defensible
State fisheries regulator
A move to digital catch reporting couldn't proceed until the data underneath a high-stakes fishery was trusted — and its quality had been publicly questioned. We weren't there to build the system; we were brought in to judge it. We audited the data quality and processing standards, set the requirements the new reporting tools had to meet, and designed the mobile interface fishers and inspectors use in the field. The foundations are ones we validated — so the data behind quota decisions can stand up to challenge.
Live · ongoing Strategic advisory