sagemaker-mlflow

What it is and what it does

This is an AWS plugin for MLflow that bridges MLflow's experiment tracking and model registry with Amazon SageMaker. It generates Signature V4 headers for outgoing requests, handles AWS IAM authentication and authorization, and enables model registration directly to the SageMaker Model Registry. The plugin works as an MLflow entry point and signs requests using credentials from the boto3 default chain (environment variables, shared config, instance role) or a custom boto3.Session you provide.

The plugin is lightweight by default, depending only on boto3 and mlflow-skinny, with an optional [full] extra for the complete MLflow package. It supports Python 3.8 through 3.14 and is actively maintained. You can inject a custom AWS session via a context manager (use_session) or set a default (set_session) without mutating environment variables, making it suitable for multi-tenant or multi-profile scenarios.

Use it for:

• Run MLflow experiment tracking on SageMaker with automatic AWS IAM authentication instead of managing separate credentials.
• Register trained models to the SageMaker Model Registry directly from MLflow without manual AWS API calls.
• Use different AWS profiles or per-tenant credentials in a shared process by injecting custom boto3 sessions.
• Integrate MLflow into SageMaker training jobs or SageMaker Pipelines with minimal configuration overhead.
• Sign MLflow requests with AWS IAM for audit trails and fine-grained access control in enterprise environments.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates MLflow with Amazon SageMaker by signing requests with AWS IAM credentials and enabling model registration to the SageMaker Model Registry.

Yes. This plugin is worth installing if you use MLflow with SageMaker and need AWS IAM-based authentication. It has low install friction, active maintenance, no known vulnerabilities, and a permissive license. The tight integration with boto3 and support for custom sessions makes it practical for both simple and multi-tenant deployments.

Install

sagemaker-mlflow on PyPI

pip install sagemaker-mlflow

uv add sagemaker-mlflow

poetry add sagemaker-mlflow

Installing sagemaker-mlflow

Before you install

Low friction installation with minimal dependencies—depends only on boto3 and mlflow-skinny by default. Active maintenance with a recent release and no known vulnerabilities.

License in practice

Licensed under Apache License 2.0 (permissive), allowing commercial use and modification with minimal restrictions.

Quickstart

pip install sagemaker-mlflow

import boto3
import mlflow
import sagemaker_mlflow

custom = boto3.Session(profile_name="my-profile")
with sagemaker_mlflow.use_session(custom):
    mlflow.MlflowClient().search_experiments(max_results=1)

Requires AWS credentials available via boto3 (environment variables, shared config, instance role, or explicit session).

Verify before relying

• Whether the plugin works with all MLflow versions listed in tox.ini or only specific ones.
• Performance overhead of SigV4 signing on high-volume MLflow tracking requests.

Package facts

Evidence: sagemaker_mlflow-0.5.0-py3-none-any.whl

Tags

More Artificial Intelligence packages