US 12,340,239 B2
Method and system for application programming interface based container service for supporting multiple machine learning applications
Alexandr Nikitin, El Sobrante, CA (US); Vaibhav Gumashta, San Francisco, CA (US); Manoj Agarwal, Cupertino, CA (US); and Swaminathan Sundaramurthy, Los Altos, CA (US)
Assigned to Salesforce, Inc., San Francisco, CA (US)
Filed by Salesforce, Inc., San Francisco, CA (US)
Filed on Jun. 2, 2021, as Appl. No. 17/337,389.
Prior Publication US 2022/0391748 A1, Dec. 8, 2022
Int. Cl. G06N 20/00 (2019.01); G06F 9/455 (2018.01)
CPC G06F 9/45558 (2013.01) [G06F 2009/45595 (2013.01)] 21 Claims
OG exemplary drawing
 
1. A method in a machine learning serving infrastructure that receives a request from a client application for a scoring, the method comprising:
receiving, at a base scorer in a scoring service container executed in the machine learning service infrastructure, at least a set of parameters from the request according to a bring your own code (BYOC) application programming interface (API), wherein the BYOC API provides an interface between the machine learning service infrastructure and scoring service containers;
calling, by the base scorer based on a model identifier, a model loader of an application specific scorer in the scoring service container, wherein the application specific scorer was developed by a machine learning model developer;
receiving, by the base scorer, a model object from the model loader, wherein the model object is a machine learning model;
passing, by the base scorer, at least the set of parameters from the request to a scoring function that is also of the application specific scorer and that is to generate the scoring responsive to application of the set of parameters to the machine learning model;
receiving, by the base scorer, the scoring from the application specific scorer; and
returning, by the base scorer and according to the BYOC API, the scoring to the machine learning serving infrastructure for delivery to the client application.