Govern models with MLflow and Amazon SageMaker AI Model Registry sync: Part 2
Governing models across accounts is the next step after automatic model registration. This post extends managed MLflow and Amazon SageMaker AI Model Registry sync to two cross-account governance topologies: a hub-and-spoke pattern that centralizes governance with AWS RAM, and a hybrid pattern that keeps development accounts isolated.
In the hub-and-spoke governance topology, a central governance account hosts a Model Registry and MLflow app. Development accounts, or spoke accounts, connect to this shared registry. The administrator sets up cross-account access by creating an AWS Resource Access Manager (AWS RAM) resource share for the MLflow app. The spoke account accepts the invitation and grants access to the shared artifact bucket in the hub account via an S3 bucket policy.
The data scientist registers the model in the spoke account, and MLflow automatically creates a Model Package Group and version in the hub account. The governance officer validates the model and approves it centrally in the hub. Once approved, CI/CD pipelines deploy the model to an endpoint in the spoke account.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
