Showcase

cat ./showcase.txt

Our mission is to forecast the coupled world so that free societies can foresee risk and secure a bright future. Each project below is a step towards that end.

bash ./deeploom.sh

DEEPLOOM: Planetary-Scale Forecasting Model

DeepLoom is a heterogeneous graph neural network that unifies spatial nodes (planetary physics) with entity nodes (infrastructure and economics). It's designed to learn the interactions classic models miss, producing auditable, multi-domain forecasts that anticipate cascading interactions.

status: In development

bash ./weavecast.sh

WEAVECAST: Forecast Delivery Platform

The WeaveCast Platform is the gateway to DeepLoom insights: one auth, uniform interaction schemas, and forecast layers for all usecases. Receive auditable, multi-domain outputs or inject your own data and integrate new domains.

status: In development

explore weavecast →

bash ./minecart.sh

MINECART: Freight bid decision support

MineCart helps freight teams price loads with clear win, margin, and destination context, forecasting freight markets to score loads in real time. This decision-support approach directly informs how we'll model market risk in DeepLoom.

status: Live

next: Grow adoption and expand market coverage

partner: Werner Enterprises

explore minecart →

bash ./mjolnur.sh

MJOLNuR: Meteorological Joint Observations for Live Nuclear Reconnaissance

MJOLNuR Phase I is a Small Business Innovation Research (SBIR) award for developing a multi-source satellite data fusion and real-time plume modeling and forecasting system. The digital twin technology for nuclear plumes directly informs how we'll model atmospheric phenomena in DeepLoom.

status:DTRA SBIR Phase I complete, Phase II proposal submitted

next: Advance to follow-on work

partner: Defense Threat Reduction Agency