Venture build · RedZone Defense
From drone flight to
risk score in one platform.
We partnered with RedZone Defense to build and launch a wildfire risk assessment platform for properties in Colorado's wildland-urban interface. We built all of it: the computer vision model, the scoring model, the platform, and the risk reports.
The challenge
Turn a property walk-through into data.
A wildfire risk assessment usually depends on one person's site visit and judgment. RedZone wanted assessments that are repeatable, backed by imagery and public wildfire data, and delivered as a report a homeowner can act on, with a path to the contractors who can do the work.
What we built
Four systems, one pipeline.
- 01
Computer vision model
Multispectral drone capture is mapped band by band and run through an NDVI pipeline to measure vegetation health. Segmentation models identify vegetation, fuels, and structures around the home, with polygon tools for reviewing and correcting results.
- 02
Risk scoring model
A scoring engine combines the CV output with six external datasets: LANDFIRE, Colorado State Forest Service data, the Colorado Wildfire Risk Assessment, USFS forest health, wildland-urban interface mapping, and fuel models. The result is one property-level risk score with the factors behind it.
- 03
Platform
A four-step assessment workflow handles imagery upload (including GeoTIFFs), analysis, and review. Role-based dashboards serve homeowners, contractors, and admins, and Stripe Connect holds contractor payments in escrow until work is done.
- 04
Risk reports
Each assessment produces an automated PDF report that explains the score, shows what was found on the property, and lays out recommended mitigation in plain language.
Stack
Built to run in production.
Next step
Have something worth building?
Tell us what you're working on and we'll take it from there.