From Predictions to Decisions: RasterFlow in Production

Name: From Predictions to Decisions: RasterFlow in Production
Start: 2026-11-11T08:00:00-08:00

The USDA Forest Service and Miraterra share how they run Earth imagery inference in production with RasterFlow.

DateWednesday, November 11, 2026

Time8:00 AM PT / 11:00 AM ET

WhereOnline · Live Q&A

Save your seat →

About RasterFlow

RasterFlow is a serverless computer vision engine for satellite and aerial imagery. It builds mosaics, runs built-in or custom models, and converts predictions into vector features, all in your own S3 buckets.

Who you’ll hear from

FireCon wildfire containment predictions across the western United States, generated with RasterFlow

USDA Forest Service

USDA Forest Service delivers wildfire containment predictions to its Wildfire Risk Management Division at an operational cadence. Each FireCon model run finishes in hours, so the team can run it several times a day.

Agricultural field boundaries predicted from Sentinel-2 imagery with RasterFlow

Miraterra

Miraterra predicts agricultural field boundaries across the U.S. Midwest and Canada’s Prairie Provinces, then joins microbial samples and geospatial embeddings to each field.

What you’ll learn

How each team moved its imagery workload into production
How they estimate cost and scale from a test area to full coverage
Live Q&A with both teams

Speakers

Christopher D. O’Connor, PhDResearch Ecologist, USDA Forest Service

Dave McCaffreyDirector, Strategic Funding, Miraterra

Ben PrudenHead of GTM, Wherobots