Description

Introduction

MLOps (Machine Learning Operations) bridges the gap between data science and production engineering by streamlining the ML lifecycle—from data preparation to model deployment and monitoring. Vertex AI provides native MLOps capabilities to automate, manage, and scale ML pipelines on Google Cloud. This course introduces the foundational concepts of MLOps and demonstrates how Vertex AI supports operationalizing ML workflows effectively.

Prerequisites

• Familiarity with basic ML concepts (training, inference, models)

• Basic knowledge of Google Cloud Platform (GCP) services

• Optional: Python and YAML syntax familiarity

• Vertex AI, Cloud Storage, and IAM permissions set up

Table of Contents

1. Introduction to MLOps
   1.1 What is MLOps and Why It Matters
   1.2 Key Components: CI/CD, Reproducibility, Monitoring
   1.3 MLOps vs DevOps: Similarities and Differences
   1.4 Benefits and Challenges of MLOps

2. Overview of Vertex AI for MLOps
   2.1 Vertex AI Architecture
   2.2 Built-in MLOps Tools and Integrations
   2.3 Comparison with Other ML Platforms

3. Data and Feature Engineering
   3.1 Using Vertex AI Feature Store
   3.2 Managing Feature Pipelines
   3.3 Versioning and Reusability of Features

4. ML Pipelines with Vertex AI
   4.1 Introduction to Vertex Pipelines
   4.2 Creating a Pipeline using Kubeflow Pipelines (KFP)
   4.3 CI/CD for Pipelines with Cloud Build
   4.4 Reusing and Managing Pipeline Components

5. Model Training and Tuning
   5.1 Running Custom Training Jobs
   5.2 Hyperparameter Tuning with Vizier
   5.3 Using Prebuilt Containers for Training
   5.4 Storing and Registering Models

6. Model Deployment and Serving
   6.1 Creating and Managing Endpoints
   6.2 A/B Testing and Traffic Splitting
   6.3 Securing Model Endpoints
   6.4 Online vs Batch Prediction

7. Model Monitoring and Governance
   7.1 Setting Up Model Monitoring
   7.2 Drift Detection and Alerting
   7.3 Model Lineage and Audit Trails
   7.4 Responsible AI and Explainability

8. CI/CD and Automation for ML
   8.1 GitOps and Source Control Best Practices
   8.2 Automating ML Workflows with Cloud Build
   8.3 Integrating Vertex AI with Terraform or Deployment Manager

9. Real-World MLOps Use Cases
   9.1 Case Study: Fraud Detection Pipeline
   9.2 Case Study: Image Classification with Automated Retraining
   9.3 Patterns in Continuous Training and Deployment

10. Best Practices and Next Steps
    10.1 Resource Naming and Organization
    10.2 Cost Optimization for ML Pipelines
    10.3 Next Steps: Advanced MLOps, Certifications, and Projects

Vertex AI empowers teams to implement MLOps at scale with minimal overhead. By combining data engineering, model management, and deployment automation, Vertex AI helps you build production-grade ML systems that are scalable, repeatable, and governed.
This foundational training prepares you to adopt MLOps practices using real tools and real infrastructure—paving the way for enterprise-level ML workflows.