Overview

This course features Coursera Coach! A smarter way to learn with interactive, real-time conversations that help you test your knowledge, challenge assumptions, and deepen your understanding as you progress through the course. In this hands-on course, you will master managing the entire lifecycle of machine learning models on Google Cloud Platform (GCP). Starting with setting up your environment, you’ll learn about CI/CD pipelines, model deployment using Cloud Run, and automating workflows with tools like Airflow and Kubeflow. Key topics like continuous training, version control, hyperparameter tuning, and model explainability will also be covered. Using Vertex AI and GCP services, you’ll gain real-world experience with model training, batch prediction, and scaling. This course is designed for machine learning engineers, data scientists, and software engineers. Basic knowledge of machine learning concepts and Google Cloud Platform is recommended. By the end, you’ll be able to deploy, monitor, and scale ML models on GCP, making you proficient in ML Ops practices and cloud-based model management.

Syllabus

Introduction & prerequisites

In this module, we will introduce the course objectives and setup procedures, including creating a GCP trial account, configuring the gcloud CLI, and reviewing the course structure. You'll also learn about the key Google Cloud services that will be used throughout the course.

Introduction to ML Ops

In this module, we will introduce you to the fundamental concepts of ML Ops, covering its principles, components, and the critical role it plays in managing the lifecycle of machine learning models.

CI/CD using GCP CloudBuild, Artifact & Container Registry and CloudRun

In this module, we will dive into the CI/CD processes on Google Cloud, teaching you how to automate build, testing, and deployment workflows for machine learning models and applications. You’ll gain hands-on experience deploying Flask applications and setting up automated deployment pipelines.

Continuous Model Training using Cloud Composer-Airflow

In this module, we will explore continuous model training with Cloud Composer and Airflow, focusing on automating the retraining process and handling failure scenarios in machine learning workflows. You’ll also learn how to implement logging, alerting, and CI/CD for model training pipelines.

Vertex AI For Data Science & Machine Learning

In this module, we will delve into Vertex AI, Google Cloud’s end-to-end platform for machine learning. You’ll learn how to train, deploy, and manage models with Vertex AI, while also exploring automated prediction services and CI/CD integration.

Vertex AI-Kubeflow Pipelines for ML Workflow Orchestration

In this module, we will cover Kubeflow for workflow orchestration, showing how to deploy and manage machine learning pipelines using Vertex AI. You’ll implement hyperparameter tuning, train multiple models, and complete real-world assignments.

Vertex AI-Hyperparameter Tuning Jobs, Explainability AI & Model Versioning

In this module, we will explore advanced techniques for hyperparameter tuning, model explainability, and versioning in Vertex AI. You’ll gain hands-on experience deploying and evaluating models with explainability features, while also managing different versions of your models.

Generative AI on Google Cloud

In this module, we will introduce you to the world of Generative AI on Google Cloud. You will learn about Google’s PaLM 2 language model and implement various generative AI applications, deploying models on Cloud Run and exploring hands-on labs.