MLOps with Feature Store – Move models from development to production

AIEngineering

#datascience #machinelearning #featurestore

MLOps (Machine Learning Operations) is a recent term that is concerned with how to automate model training, model validation, and model deployment. MLOps can be thought of as an extension of DevOps (Software Development Operations) with the goal to unify both ML applications development and operation from ML applications, making it easier for teams to deploy better models more frequently and more efficiently. A feature store is a feature computation and storage service that enables features to be registered, discovered, and used both as part of ML pipelines as well as by online applications for model inferencing. It enables engineers to apply software engineering development principles to ML, more precisely creating a hub for feature data where people and users can discover and share features and with other users within the organization. During this talk, Jim Dowling, CEO at Logical Clocks, will discuss and demo how the Hopsworks Feature Store enables MLOps workflows for training models using features from the Feature Store, analyzing and validating models, deploying them into online model serving infrastructure, and monitoring model performance in production.

Source