SDK v0.9.x

Integrations

Inferpathio is designed to layer on top of your existing ML infrastructure. These guides walk through connecting each supported tool.

MLflow

Connect MLflow artifact runs to Inferpathio governance without changing your existing training code.

import mlflow
import inferpathio as ifp

with mlflow.start_run() as mlflow_run:
    with ifp.governance_run(model_id='fraud-detector-v2') as gov:
        # Your existing MLflow training code
        mlflow.log_param("n_estimators", 100)
        mlflow.sklearn.log_model(model, "model")
        # Link MLflow run to governance event
        gov.link_mlflow_run(mlflow_run.info.run_id)
        gov.set_policy("production-strict")

SageMaker

Connect SageMaker inference endpoints to Inferpathio drift monitoring via the SageMaker Data Capture feature.

from inferpathio.integrations import SageMakerConnector

connector = SageMakerConnector(
    endpoint_name="fraud-detector-prod",
    model_id="fraud-detector-v2",
    policy="production-strict"
)

# Enable data capture on your endpoint
connector.enable_monitoring(
    capture_percentage=100,
    s3_bucket="my-ml-data-lake"
)

GitHub Actions

Fire Inferpathio governance events from your existing GitHub Actions training workflows.

on:
  repository_dispatch:
    types: [retrain-approved]

jobs:
  retrain:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Train model
        run: python train.py
        env:
          INFERPATH_KEY: ${{ secrets.INFERPATH_KEY }}
          GOVERNANCE_RUN_ID: ${{ github.event.client_payload.run_id }}

Airflow

Use the Inferpathio Airflow operator to notify governance on DAG completion events.

from inferpathio.integrations.airflow import GovernanceCompleteOperator

# Add at end of retrain DAG
notify_governance = GovernanceCompleteOperator(
    task_id='notify_governance',
    model_id='fraud-detector-v2',
    run_id='{{ ti.xcom_pull("start_governance") }}'
)