<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Valkey on Jaehyeon Kim</title><link>https://jaehyeon.me/tags/valkey/</link><description>Recent content in Valkey on Jaehyeon Kim</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>Copyright © 2023-2026 Jaehyeon Kim. All Rights Reserved.</copyright><lastBuildDate>Mon, 23 Feb 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://jaehyeon.me/tags/valkey/index.xml" rel="self" type="application/rss+xml"/><item><title>Productionizing an Online Product Recommender using Event Driven Architecture</title><link>https://jaehyeon.me/blog/2026-02-23-productionize-recommender-with-eda/</link><pubDate>Mon, 23 Feb 2026 00:00:00 +0000</pubDate><guid>https://jaehyeon.me/blog/2026-02-23-productionize-recommender-with-eda/</guid><description><![CDATA[<p>In <a href="/blog/2026-01-29-prototype-recommender-with-python/"><strong>Part 1</strong></a>, we built a contextual bandit prototype using Python and <a href="https://github.com/fidelity/mab2rec" target="_blank" rel="noopener noreferrer"><code>Mab2Rec</code><i class="fas fa-external-link-square-alt ms-1"></i></a>. While effective for testing algorithms locally, a monolithic script cannot handle production scale. Real-world recommendation systems require low-latency inference for users and high-throughput training for model updates.</p>
<p>This post demonstrates how to decouple these concerns using an event-driven architecture with Apache Flink, Kafka, and Valkey.</p>]]></description><enclosure url="https://jaehyeon.me/blog/2026-02-23-productionize-recommender-with-eda/featured.gif" length="702786" type="image/gif"/></item></channel></rss>