Back to AI/ML

Vertex AI Pipelines

gcp

Serverless orchestration for automating, monitoring, and governing end-to-end MLOps workflows.

AI & ML
Google Cloud Platform

Serverless orchestration for automating, monitoring, and governing end-to-end MLOps workflows.

Click any section above to jump directly to it

What is Vertex AI Pipelines?

Vertex AI Pipelines is a serverless workflow orchestrator based on Kubeflow Pipelines (KFP) and TensorFlow Extended (TFX). It automates the end-to-end machine learning lifecycle without requiring you to manage underlying Kubernetes clusters. Every pipeline run automatically captures input data, parameters, model metrics, and generated artifacts in Vertex ML Metadata for complete auditability and reproducibility.

Decision Guide: When to Choose This Service

When to Use Vertex AI Pipelines

Use Vertex AI Pipelines to automate repeatable machine learning workflows: automated data extraction, model training, evaluation validation checks, conditional deployment, and continuous retraining with full lineage tracking.

When NOT to Use (Recommended Alternatives)

Do NOT use Vertex AI Pipelines for general data engineering ETL that doesn't involve machine learning (use Cloud Composer or Dataflow).

Core Architectural Pillars

Kubeflow Pipelines (KFP) SDK

Author MLOps pipelines as modular, reusable Python functions packaged into container components.

Vertex ML Metadata

Automatically tracks lineage of all artifacts (datasets, models, evaluation metrics) across pipeline runs.

Continuous Retraining

Trigger pipeline executions automatically via Eventarc or Cloud Scheduler when data drift is detected.

Serverless Execution

Pay only for the exact seconds each container component executes with zero cluster idle costs.