FastScore is an analytic deployment environment, enabling organizations to iterate and deploy more models into production faster. With an agnostic deployment environment, FastScore helps organization with heterogeneous analytic creation tools and a variety of data sources to quickly and easily turn "models" into "assets" and build competency around monetizing those assets.
Open Data Group is proud to announce the FastScore v1.2 release. New for FastScore v1.2 includes these key features and capabilities:
- Significant performance improvements from a new stream handling architecture
- Unified model language support in Engine including (R, Python, Java, PFA and PrettyPFA)
- Support for TCP, UDP, debug, and console streams (in addition to the already-supported HTTP, Kafka, and file streams)
Universal Deployment, Start to Finish
THE FASTSCORE ENGINE DEPLOYS MODELS REGARDLESS OF THEIR ORIGINAL LANGUAGE FROM START TO FINISH
- Eliminates model conversion when it’s not required AND streamlines the process when it is
- Allows rapid deployment for more robust models: data scientists no longer need to sacrifice model complexity for deployment complications
Intelligent Interaction with Analytics
EVER WONDER IF THAT NEW RECURRENT NEURAL NET MODEL PROVIDES TRUE LIFT TO YOUR BUSINESS? HAS THE IT TEAM STRUGGLED TO UNDERSTAND WHY A LARGE RANDOM FORREST MODEL NEEDS SO MUCH MEMORY TO RUN? TRANSLATE ANALYTICS INTO VALUABLE INSIGHTS FOR YOUR COMPANY AND CUSTOMERS WITH SMART INSTRUMENTATION AND TOOLING. The Fastscore engine deploys models regardless of their original language from start to finsih
- Scores as quickly as your data streams
- Measures the model's performance as it scores, allowing you to react faster
- Monitor model health including instrumenting CPU, memory usage and more.
Fit to your Business
Fastscore forms to the needs of the business: scoring on real time, telematic data? No problem. Batch scoring of credit risk profiles? Covered. Developing a strategy that will increase analytic use over time? That's what fastscore was built for.
- Supports on-going analytic deployments with one time integration to the IT stack
- Allows for running models at scale without harm to deployment speed
- Deploys from any platform - in the cloud, on premise, or on a single laptop