Feature Engineering Course Syllabus

Overview: This course provides a hands-on introduction to feature engineering using Google Cloud Platform and TensorFlow tools. You'll learn to transform raw data into production-ready features, explore transformations like bucketing and feature crosses, and integrate with MLOps workflows. The course spans approximately 8.3 hours, combining theory, labs, and practical exercises to build foundational skills in modern ML pipelines.

Module 1: Introduction to Vertex AI Feature Store

Estimated time: 0.8 hours

• Understand the purpose and benefits of a feature store
• Explore core components of Vertex AI Feature Store
• Learn setup and configuration basics
• Review use cases for centralized feature management

Module 2: Raw Data to Features

Estimated time: 1 hour

• Identify sources of raw data suitable for ML
• Define criteria for feature quality and relevance
• Establish best practices for feature selection
• Map raw data to potential ML features

Module 3: Feature Engineering Basics

Estimated time: 4 hours

• Compare ML and statistical approaches to feature engineering
• Apply feature transformations using BigQuery ML and Keras
• Implement feature crosses and bucketing techniques
• Use tf.Transform for scalable preprocessing

Module 4: Advanced Feature Engineering & MLOps

Estimated time: 2 hours

• Apply advanced feature transformations in TensorFlow
• Handle metadata and versioning in feature pipelines
• Integrate feature engineering with MLOps workflows

Module 5: Course Conclusion

Estimated time: 0.5 hours

• Review key feature engineering best practices
• Summarize tools and platforms covered
• Reflect on end-to-end feature design and production integration

Prerequisites

• Familiarity with machine learning concepts and workflows
• Basic knowledge of TensorFlow and Keras
• Experience with Google Cloud Platform (GCP) tools

What You'll Be Able to Do After

• Use Vertex AI Feature Store to manage and serve features
• Transform raw data into ML-ready features using BigQuery ML and TensorFlow
• Apply feature crosses, bucketing, and tf.Transform in practice
• Integrate feature pipelines into MLOps systems
• Follow best practices for preprocessing and model accuracy enhancement