dSystems.ai Deep Learning models run from a colocated GPU server in New Jersey. Forecasts are available via a low-latency API.