Does it help with model monitoring and drift?

Absolutely, the skill provides strategies for detecting model performance drift and setting up production observability for deployed models.

What tools does this skill support?

It provides implementation patterns for major orchestration tools like Apache Airflow, Dagster, and Kubeflow, as well as experiment tracking tools like MLflow and Weights & Biases.

Can I use this for real-time ML pipelines?

Yes, it includes guidance and patterns for both batch training pipelines and real-time stream processing for feature engineering.

Is this skill compatible with major cloud providers?

Yes, it includes integration points and deployment patterns for AWS SageMaker, Google Vertex AI, and Azure ML infrastructure.

How does this help with MLOps best practices?

It standardizes the transition from experimental notebooks to production-ready automated pipelines, emphasizing versioning, idempotency, and modularity.

ML Pipeline Workflow

Name: ML Pipeline Workflow
Author: Microck

byMicrock

•

データサイエンスとML

Builds end-to-end MLOps pipelines from data preparation through model training, validation, and production deployment.

概要

This skill provides comprehensive guidance for orchestrating production-grade machine learning pipelines, covering the entire lifecycle from data ingestion to monitoring. It offers structured patterns for DAG orchestration using tools like Airflow and Kubeflow, integrates experiment tracking with MLflow, and implements robust deployment strategies like canary or blue-green releases. Whether building a pipeline from scratch or automating existing workflows, this skill ensures reproducibility, scalability, and observability in ML systems.

主な機能

Integration with industry-standard MLOps tools (Airflow, MLflow, Vertex AI)
81 GitHub stars
Automated data validation and feature engineering patterns
End-to-end DAG orchestration for ML workflows
Production deployment automation and rollback strategies
Comprehensive model validation and A/B testing frameworks

ユースケース

Implementing automated CI/CD for machine learning (MLOps)
Designing modular data preparation and feature engineering pipelines
Creating reproducible training workflows for production models

概要

主な機能

Integration with industry-standard MLOps tools (Airflow, MLflow, Vertex AI)
81 GitHub stars
Automated data validation and feature engineering patterns
End-to-end DAG orchestration for ML workflows
Production deployment automation and rollback strategies
Comprehensive model validation and A/B testing frameworks

ユースケース

Implementing automated CI/CD for machine learning (MLOps)
Designing modular data preparation and feature engineering pipelines
Creating reproducible training workflows for production models