We apply rigorous mathematical principles and proprietary techniques to provide accurate solutions to complex data problems. With a core focus on both accuracy and confidence, our solutions are developed from the ground up to track data and algorithmic uncertainty while maintaining the data provenance needed to make informed, confident decisions.

RootShape is a machine learning tool for the segmentation of plant roots from the background soil in minirhizotron images. Using Topological Data Analysis techniques and an active learning framework, RootShape provides rapid, accurate, and consistent image analysis with minimal user guidance, significantly reducing the cost and time required for root segmentation.

Localized Wind Modeling combines low resolution wind forecasts with topographic information to improve spatial accuracy and model forecast uncertainty.