Engine is the core component of FastScore that executes and scores a machine learning model. Each engine is easily customized to the specific execution environment that the individual model requires and is agnostic to the model language. Engine consumes and produces data through multiple methods of data transportation (REST, Kafka, ODBC, TCP, …). The Engine, containing an individual execution model, can be scaled for performance and monitored, and managed alongside any other microservice using your orchestration technology of choice.
Allows for multiple input and output data connections
Easy configuration for any model environment
Supports multiple data encodings
Inter-engine scalability to easily create additional model instances
Supports multiple data transport connectors
Infrastructure agnostic that can run anywhere docker is supported