Vertex AI for Machine Learning Practitioners

(1 day)

 

This instructor-led one-day course is designed for engineers and data scientists familiar with machine learning models who want to become proficient in using Vertex AI for custom model workflows. This practical, hands-on course will provide you with a deep dive into the core functionalities of Vertex AI, enabling you to effectively leverage its tools and capabilities for your ML projects.

Course Objectives

• Understand the key components of Vertex AI and how they work together to support your ML workflows.
• Configure and launch Vertex AI Custom Training and Hyperparameter Tuning Jobs to optimize model performance.
• Organize and version your models using Vertex AI Model Registry for easy access and tracking.
• Configure serving clusters and deploy models for online predictions with Vertex AI Endpoints.
• Operationalize and orchestrate end-to-end ML workflows with Vertex AI Pipelines for increased efficiency and scalability.
• Configure and set up monitoring on deployed models

Audience

• Machine Learning Engineers, Data Scientists

Prerequisites

• Experience building and training custom ML models. Familiar with Docker.

---

Course Outline

 

Module 1: Training, Tuning, and Deploying Models on Vertex AI

• Understand Containerized Training Applications
• Understand Vertex AI Custom Training and Tuning Jobs
• Understand how to track and version your trained models in Vertex AI Model Registry
• Understand Online Deployment with Vertex AI Endpoints

Module 2: Orchestrating End-to-End Workflows with Vertex AI Endpoints

• Understand Kubeflow
• Understand pre-built and lightweight Python components
• Understand how to compile and execute pipelines on Vertex AI

Module 3: Model Monitoring on Vertex AI

• Understand Feature Drift and Skew
• Understand Model Monitoring for models deployed to Vertex AI Endpoints