Feature store
A feature store is a system for managing and serving machine-learning features for use in model training and inference.
Machine Learning Platform Engineering
A feature store is a system that centrally stores, manages, and serves curated machine-learning features for training and inference.
- How it is used
- Positions a feature store between feature preparation and model use to share curated features for training and inference.
- How deep it goes
- Explains Feast’s registry, offline and online stores, materialization, and point-in-time retrieval; leaves retraining and feature-update automation to pipelines.
- Examples
- Uses timestamped user features for point-in-time retrieval and a tabular movie recommender to motivate feature-store tooling.
- What you can do after
- Apply it via Feast to register and share features, retrieve historical feature sets, and serve features from an online store.
Read how Machine Learning Platform Engineering explains feature store →
LLMs in Production
A feature store is a system or abstraction for managing, documenting, versioning, governing, and serving machine-learning features for training and production.
- How it is used
- Provides a governed source of features for collaboration, training, and production inference.
- How deep it goes
- Explains the abstraction and production trade-offs, then demonstrates Feast setup, a feature view, materialization, and an online query.
- Examples
- Uses a rolling request average and a Feast question-and-answer dataset with Parquet-backed features and an online query.
- What you can do after
- Apply it via Feast to define, register, materialize, and query a feature view for inference.
Machine Learning System Design
An infrastructure system for storing, managing, and serving machine-learning features.
- How it is used
- Centralizes feature calculation, storage, testing, documentation, and monitoring so ML teams can reuse features across pipelines.
- How deep it goes
- Works through benefits and design questions around read latency, update schedules, and precomputation, without specifying a full implementation.
- Examples
- Uses duplicated feature implementations, daily order processing, and a diagram contrasting online and batch feature writes.
- What you can do after
- Evaluate/understand it by weighing reuse against management effort and checking read latency, write load, and refresh schedules.
Read how Machine Learning System Design explains feature store →
Delivering Machine Learning Projects: From design to deployment
A feature store is a system for storing, managing, and serving machine learning features for model development and inference.
- How it is used
- Supports feature reuse and consistent data use while fitting into the ML team’s development and production infrastructure.
- How deep it goes
- Explains organizational value and hosting considerations, but stops short of specifying storage schemas, APIs, or implementation mechanics.
- Examples
- Uses a reusable sensor-position representation as an example of domain insight preserved for future projects.
- What you can do after
- Build it from scratch by planning hosting, access, supporting infrastructure, and reusable feature capture for future projects.
Read how Delivering Machine Learning Projects: From design to deployment explains feature store →
Financial AI in Practice
A centralized repository for storing, managing, and serving validated features for machine-learning workflows.
- How it is used
- Centralizes reusable financial-model features for consistent batch, training, inference, and low-latency scoring.
- How deep it goes
- Explains reuse, versioning, serving, lineage, and pipeline integration, but specifies no fixed implementation or stack.
- Examples
- Uses rolling_30d_spend and bureau flags for credit workflows, and transaction, login, and behavior events for fraud scoring.
- What you can do after
- Operate it with Airflow DAGs to publish daily aggregates and flag incomplete bureau feeds.
Related concepts
- inputuses · An input is a value, information, or signal supplied to a system, computation, or interface for processing or to influence its behavior.
- training-serving skewaddresses · Training-serving skew is a mismatch between the data, feature representations, or computations used during model training and those used to generate predictions.
- Apache Airflowcontrasts with · Apache Airflow is an open-source platform for authoring, scheduling, and monitoring programmatically defined workflows, commonly represented as directed acyclic graphs.
- machine learninguses · Machine learning is a field of computational methods in which systems learn patterns or relationships from data to make predictions, decisions, or other outputs.
- independent variableuses · An independent variable is an input or explanatory attribute whose values are used to model, explain, or predict variation in an outcome.
- source of truthaddresses · A source of truth is the authoritative source or representation from which the accepted value or state of information is determined.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.