<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>AI Infrastructure on Sparse Notes</title>
    <link>https://sparsenotes.com/categories/ai-infrastructure/</link>
    <description>Recent content in AI Infrastructure on Sparse Notes</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en</language>
    <copyright>© 2026 Pradithya Aria Pura</copyright>
    <lastBuildDate>Sat, 31 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://sparsenotes.com/categories/ai-infrastructure/index.xml" rel="self" type="application/rss+xml" />
    
    <item>
      <title>How Revolut Cut AI Coding Costs 90% With Self-Hosted Open Models</title>
      <link>https://sparsenotes.com/posts/2026/10/revolut-self-hosted-open-models-david-abitbol/</link>
      <pubDate>Sat, 31 Oct 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/10/revolut-self-hosted-open-models-david-abitbol/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Mixture-of-Kittens: MoE Megakernel for NVL72s</title>
      <link>https://sparsenotes.com/posts/2026/10/papers/mixture-of-kittens-moe-megakernel-nvl72/</link>
      <pubDate>Sat, 03 Oct 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/10/papers/mixture-of-kittens-moe-megakernel-nvl72/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>From LLM Inference to Agentic Workloads: Characterization and Implications for Serving Systems</title>
      <link>https://sparsenotes.com/posts/2026/08/papers/agentsysbench-agentic-serving-workloads/</link>
      <pubDate>Sat, 22 Aug 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/08/papers/agentsysbench-agentic-serving-workloads/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Who Should Own the Expert Cache? Kernel-Managed Tiering for Trillion-Parameter MoE Inference</title>
      <link>https://sparsenotes.com/posts/2026/08/papers/kernel-managed-expert-cache-moe-inference/</link>
      <pubDate>Sat, 15 Aug 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/08/papers/kernel-managed-expert-cache-moe-inference/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Architectural Implications of Agentic AI Workflows</title>
      <link>https://sparsenotes.com/posts/2026/08/papers/architectural-implications-agentic-ai-workflows/</link>
      <pubDate>Sat, 08 Aug 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/08/papers/architectural-implications-agentic-ai-workflows/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>On Topology&#39;s Role in ML Training Performance</title>
      <link>https://sparsenotes.com/posts/2026/08/papers/topology-role-ml-training-performance/</link>
      <pubDate>Sat, 08 Aug 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/08/papers/topology-role-ml-training-performance/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Incast-Free MoE Rate-Based Scheduling</title>
      <link>https://sparsenotes.com/posts/2026/08/papers/incast-free-moe-rate-scheduling/</link>
      <pubDate>Sat, 01 Aug 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/08/papers/incast-free-moe-rate-scheduling/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context</title>
      <link>https://sparsenotes.com/posts/2026/07/papers/windowed-mtp-draft-kv-tax/</link>
      <pubDate>Sat, 25 Jul 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/07/papers/windowed-mtp-draft-kv-tax/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Who Needs DRAM? We Have Fiber</title>
      <link>https://sparsenotes.com/posts/2026/07/papers/fiber-memory-2607.08407/</link>
      <pubDate>Sat, 11 Jul 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/07/papers/fiber-memory-2607.08407/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>ARGUS: Production-Scale Tracing and Performance Diagnosis for 10,000&#43; GPU Clusters</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-argus-gpu-tracing/</link>
      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/2026-06-20-argus-gpu-tracing/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>AgileOS: A GPU Operating System Layer for Protected CUDA Services</title>
      <link>https://sparsenotes.com/posts/2026/06/papers/2026-06-13-agileos-gpu-operating-system/</link>
      <pubDate>Sat, 13 Jun 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/06/papers/2026-06-13-agileos-gpu-operating-system/</guid>
      <description></description>
      
    </item>
    
    <item>
      <title>Inside Jane Street&#39;s GP300 Training Data Center</title>
      <link>https://sparsenotes.com/posts/2026/05/2026-05-23-jane-street-ai-datacenter/</link>
      <pubDate>Sat, 23 May 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/05/2026-05-23-jane-street-ai-datacenter/</guid>
      <description></description>
      
    </item>
    
  </channel>
</rss>
