PropTech startup
From site photos to finished inspection reports
3D models, a custom machine learning model and an iOS app for property management.
Services
Strategy, technology
Industry
Proptech
Role
Founder
The problem
A building inspection produces hundreds of photos, notes and measurements. Turning them into a report takes hours, and two inspectors can describe the same defect in different ways. I founded a PropTech startup to move that work onto the job site and make it more consistent, using 3D capture, machine learning and a simple app.
What I did
I led the company from the first idea to working product. I managed a team of five developers and the partners we worked with, and I handled strategy, fundraising, legal and accounting alongside the product.
3D capture. I used photogrammetry to turn site photos into measurable 3D models of buildings.
A platform for large models. I converted the models into 3D Tiles, the open standard for streaming large 3D models, so they can be viewed in a browser without specialist software. A web platform sits on top.
A custom machine learning model. We collected and labelled more than 15,000 images across 10 defect classes, then trained a model to find defects in them. The early version detects five classes.
Inspector Baseline. An iOS app with a web companion for the day-to-day work: photos stamped with location and time, voice notes on site, and written defect descriptions drafted for the inspector. The inspector reviews and edits every description before it goes in a report.
Customer trials. We ran three trials on real inspections. They showed product-market fit.
Results
3
Customer trials
Top 3
Finalist, SXSW Sydney 2025
15,000+
Labelled images across 10 defect classes
5
Developers managed
Part of Blackbird Giants cohort 10. Won Fishburners pitch competitions. Featured in Forbes Australia.
Where it is now
The platform, the model and the app were built to be reused, and Inspector Baseline now continues as a Novara product. Feel free to contact me for further details.