These tutorials cover common FeatureStore scenarios — from building a recommendation system to generating large language model (LLM) embeddings. Each tutorial focuses on a specific ML workflow stage so you can start from the one that matches your current task.
Use FeatureStore to build a recommendation system
Covers the full lifecycle of a FeatureStore-backed recommendation system using feature tables:
Create and register a project in FeatureStore
Define and store features for training
Publish the trained model for online serving
Start here if you are building a recommendation system and want to understand how FeatureStore fits into the end-to-end workflow.
Best practice for FeatureStore
Shows how to manage features in a recommendation system using only the FeatureStore SDK, without depending on other Alibaba Cloud services. Use this tutorial if you want a self-contained, SDK-only approach to feature management.
Build a recommendation system by using the FeatureStore SDK for Python
This topic describes how to use the FeatureStore SDK for Python to build and publish a recommendation system.
Use AutoFE in FeatureStore
Shows how to use Automatic Feature Engineering (AutoFE) to discover and generate features automatically instead of engineering them manually:
Run AutoFE to generate features from raw data
Apply the AutoFE pipeline model to transform features in training and test datasets
Improve model performance for machine learning and deep learning tasks
Use this tutorial if you want to automate feature generation rather than writing feature transformations by hand.
Generate LLM embeddings with FeatureStore
Shows how FeatureStore fits into an LLM embedding workflow, covering both offline batch embedding and real-time online embedding.