Overview

Introduction

This tutorial shows you how to export FeatureByte features as SQL code that runs independently in your data warehouse. You can take features discovered through FeatureByte Feature Ideation and deploy them using your existing SQL infrastructure.

You might want to export SQL when you need to:

• Compute features in your existing data pipelines
• Use your own scheduling tools (Airflow, dbt, Databricks Jobs)
• Compute features directly in your warehouse for cost or compliance reasons
• Operationalize FeatureByte discoveries without ongoing FeatureByte infrastructure

How It Works

The process is straightforward:

1. Use FeatureByte to ideate and validate features
2. Export those features as SQL templates with placeholders
3. Fill in the placeholders for your specific environment
4. Schedule the SQL to run using your preferred tools

Tutorial Structure

1. SQL Export
2. Schedule SQL Execution
3. End-to-End ML Workflow

Getting Started

Begin with SQL Export to see how the export functionality works.