<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Fault Tolerance on Rik Kisnah - Blog</title><link>https://www.rik-kisnah.ai/tags/fault-tolerance/</link><description>Recent content in Fault Tolerance on Rik Kisnah - Blog</description><generator>Hugo</generator><language>en</language><lastBuildDate>Tue, 12 May 2026 09:00:00 -0700</lastBuildDate><atom:link href="https://www.rik-kisnah.ai/tags/fault-tolerance/feed.xml" rel="self" type="application/rss+xml"/><item><title>Design LLM Training Infrastructure</title><link>https://www.rik-kisnah.ai/teach/systems/design-llm-training-infrastructure/</link><pubDate>Tue, 12 May 2026 09:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/teach/systems/design-llm-training-infrastructure/</guid><description>Twenty thousand GPUs, jobs that need four thousand of them at once, and hardware that fails every nine hours. The whole design falls out of one number: how often something breaks, and what that costs you.</description></item></channel></rss>