Machine Learning Operations with Vertex AI: Model Evaluation

Overview

This course equips machine learning practitioners with the essential tools, techniques, and best practices for evaluating both generative and predictive AI models. Model evaluation is a critical discipline for ensuring that ML systems deliver reliable, accurate, and high-performing results in production. Participants will gain a deep understanding of various evaluation metrics, methodologies, and their appropriate application across different model types and tasks. The course will emphasize the unique challenges posed by generative AI models and provide strategies for tackling them effectively. By leveraging Google Cloud's Vertex AI platform, participants will learn how to implement robust evaluation processes for model selection, optimization, and continuous monitoring.

Syllabus

Welcome to the Machine Learning Operations (MLOps) with Vertex AI: Model Evaluation

This module covers the course objectives and provides an overview of the course structure.

Introduction to Model Evaluation

This module introduces model evaluation challenges and solution offerings by Vertex AI.

Model Evaluation for Generative AI

This module describes challenges of evaluating the Generative AI tasks and best practices to overcome these challenges. The module also covers the different types of model evaluation services available in Vertex AI and then introduces Vertex AI Automatic Metrics, Automatic Side by Side and Safety Bias evaluation services.