Data streaming pipeline for compute mapping systems and applications

A system, comprising:

• one or more processing units to:
• receive a request to perform a compute operation associated with information stored in a remote data source;
• extract, based on one or more processes corresponding to a format of data retrieved from the remote data source, one or more components of the data retrieved from the remote data source;
• select, from a plurality of models, a pre-trained machine learning model based, at least in part, on the request;
• provide, to a compute instance, the pre-trained machine learning model;
• provide, to the compute instance, one or more configuration files having parameters for execution of the pre-trained machine learning model;
• queue the one or more components;
• map, from a queuing location, the one or more components directly to a layer of the pre-trained machine learning model;
• and
• cause the pre-trained machine learning model to execute the compute operation using at least the one or more components and the one or more configuration files without copying the one or more components to the compute instance.

A computer-implemented method, comprising:

• receiving a request to perform a compute operation associated with information stored in a remote data source;
• extracting, based on one or more processes corresponding to a format of data retrieved from the remote data source, one or more components of the data retrieved from the remote data source;
• selecting, from a plurality of models, a pre-trained machine learning model based, at least in part, on the request;
• providing, to a compute instance, the pre-trained machine learning model;
• providing, to the compute instance, one or more configuration files having parameters for execution of the pre-trained machine learning model;
• queuing the one or more components;
• mapping, from a queuing location, the one or more components directly to a layer of the pre-trained machine learning model;
• and
• causing the pre-trained machine learning model to execute the compute operation using at least the one or more components and the one or more configuration files without copying the one or more components to the compute instance.