Engelhart Commodities Trading Partners needed a dependable way to organize nine petabytes of disparate images and apply machine-learning models at scale so that outputs could feed pre-trade decision workflows. The team required an environment that accommodated both Python and R and supported a repeatable path from research to production.
The image corpus was vast and heterogeneous, which demanded a configurable approach across images, Python, and R. The platform had to meet performance requirements for large-scale analysis while allowing the team to move quickly from exploratory work to consistent, production-grade results without losing clarity on what was run and when.
NorthGravity delivered a native, scalable production pipeline on AWS that combined image analytics with Python and R workflows. The platform enabled rapid setup and execution of model runs and produced pre-trade prescriptive outputs in less than one month. This approach gave ECTP a clear and repeatable way to test, validate, and iterate, using the same environment for research and production.
ECTP accelerated time to insight and established a reliable path from exploratory analysis to at-scale delivery. Pre-trade teams gained timely, consistent outputs, and the organization confirmed that a highly configurable platform – covering images, Python, and R – was essential to performance and delivery at scale.
