---
title: "Discussion on \"Feature Backfilling: How to Generate Historical Features for Model Retraining\" | Hashnode"
description: "Discussion on \"Feature Backfilling: How to Generate Historical Features for Model Retraining\". Reconstructing historical feature values from raw events using the same SQL logic as your live pipeline — eliminating training-serving skew."
url: "https://hashnode.com/posts/feature-backfilling-time-travel-streaming-database/69d452ead95c5055dfaebdee"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Feature Backfilling: How to Generate Historical Features for Model Retraining](https://risingwave.com/blog/feature-backfilling-time-travel-streaming-database?utm_source=hashnode&utm_medium=feed)

Reconstructing historical feature values from raw events using the same SQL logic as your live pipeline — eliminating training-serving skew.

Apr 7 · 11 min read · risingwave.com  
Read the article

-   [#backfill](https://hashnode.com/tag/backfill)
-   [#feature-store](https://hashnode.com/tag/feature-store)
-   [#mlops](https://hashnode.com/tag/mlops)
-   [#model-training](https://hashnode.com/tag/model-training)
-   [#streaming-database](https://hashnode.com/tag/streaming-database)

## Written by

[RisingWave Labs](https://hashnode.com/@RisingWaveLabs)

@RisingWaveLabs · Bay Area, California

  
[View profile](https://hashnode.com/@RisingWaveLabs)

RisingWave is an open-source real-time event streaming platform.

RisingWave is a real-time event streaming platform designed to offer the simplest and most cost-effective way to process, analyze, and manage real-time event data — with built-in support for the Apache Iceberg™ open table format. It provides both a Postgres-compatible SQL interface and a DataFrame-style Python interface. RisingWave can ingest millions of events per second, continuously join and analyze live streams with historical data, serve ad-hoc queries at low latency, and manage data reliably in Apache Iceberg™ tables.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
