<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Versioning on Rik Kisnah - Blog</title><link>https://www.rik-kisnah.ai/tags/versioning/</link><description>Recent content in Versioning on Rik Kisnah - Blog</description><generator>Hugo</generator><language>en</language><lastBuildDate>Tue, 20 Jan 2026 09:00:00 -0800</lastBuildDate><atom:link href="https://www.rik-kisnah.ai/tags/versioning/feed.xml" rel="self" type="application/rss+xml"/><item><title>Design a Model Serving Platform with Versions and A/B Tests</title><link>https://www.rik-kisnah.ai/teach/gpu-ai/design-a-model-serving-platform/</link><pubDate>Tue, 20 Jan 2026 09:00:00 -0800</pubDate><guid>https://www.rik-kisnah.ai/teach/gpu-ai/design-a-model-serving-platform/</guid><description>A hundred models from a hundred megabytes to a hundred gigabytes, clients on every continent, and product teams who want to ship version eight to ten percent of traffic tonight. Design the platform that stores the versions, gets them onto GPUs, splits the traffic, and never loads a model while a client is waiting.</description></item><item><title>Design a Prompt Playground</title><link>https://www.rik-kisnah.ai/teach/systems/design-a-prompt-playground/</link><pubDate>Tue, 09 Dec 2025 09:00:00 -0800</pubDate><guid>https://www.rik-kisnah.ai/teach/systems/design-a-prompt-playground/</guid><description>A million users, ten thousand a day, under one request a second, and prompts that can reach ten megabytes. The interviewer opens a Google Doc, sketches client, server and database, and says: no architecture diagram. The whole design is the blob path and the sharing semantics, and the whole test is whether you can keep the scope small.</description></item></channel></rss>