Operating Any Model
Any model, any language, deployed. Period.
Rather than taking over the world, let’s start with one model.
Data Science. Data Engineering. DevOps. Supported.
We build for value, and integrate to existing capabilities.
“I can just wrap my model with a script”.... It’s been said before. But building one-off scripts, by hand, for a transformative machine learning model is not the path to success for the large enterprise. Both modeling and IT teams need a standard, scalable, routine way to prepare, test and deploy models into production applications.
The FastScore Engine eliminates the need for custom built, “by-hand” approaches to model deployment. Delivered as a Docker image, Engine is built to enable robust model operations for the large enterprise. Our customers have shown Engine allows modeling, data engineering and IT teams to align to a common, flexible approach for deploying their critical machine learning assets. Via it’s core abstractions, Engine provides a path to rapid deployment, iteration and scale for any application that drives the lines of business.
Cloud ready, but not required
Enable batch, on-demand and streaming applications
Multiple performance scaling options
Supports models with state
Detailed logging and performance data