<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Meet Modi — Blog</title>
    <link>https://www.modimeet.com/blog</link>
    <description>Engineering write-ups by Meet Modi on web internals, backend systems, and the things learned building software.</description>
    <language>en</language>
    <lastBuildDate>Sun, 15 Feb 2026 00:00:00 GMT</lastBuildDate>
    <atom:link href="https://www.modimeet.com/feed.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>How ad-blockers actually work, and why they don&apos;t speed up the page</title>
      <link>https://www.modimeet.com/blog/how-ad-blockers-actually-work</link>
      <guid isPermaLink="true">https://www.modimeet.com/blog/how-ad-blockers-actually-work</guid>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <description>I thought ad-blockers stopped ads from downloading. Turns out half the time, they don&apos;t. The download already happened. The block is what you see.</description>
      <category>Web</category>
      <category>Browser</category>
      <category>Networking</category>
    </item>
    <item>
      <title>What I learned co-writing an applied-ML paper on cerebral aneurysms</title>
      <link>https://www.modimeet.com/blog/ml-surrogates-for-aneurysms</link>
      <guid isPermaLink="true">https://www.modimeet.com/blog/ml-surrogates-for-aneurysms</guid>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <description>We replaced an expensive CFD simulation with a neural network surrogate. The trick wasn&apos;t the network. It was throwing 99% of the dimensions away first.</description>
      <category>Applied ML</category>
      <category>Research</category>
      <category>Numerical Methods</category>
    </item>
    <item>
      <title>Atomic Redis counters and the race condition that took me three tries</title>
      <link>https://www.modimeet.com/blog/atomic-redis-counters</link>
      <guid isPermaLink="true">https://www.modimeet.com/blog/atomic-redis-counters</guid>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <description>I had a rate limiter. It capped most users at 100 emails per hour. A few got to send 105. The bug: INCR was atomic. My code wasn&apos;t.</description>
      <category>Redis</category>
      <category>Concurrency</category>
      <category>Backend</category>
    </item>
    <item>
      <title>Why I queue emails instead of sending them</title>
      <link>https://www.modimeet.com/blog/why-i-queue-emails</link>
      <guid isPermaLink="true">https://www.modimeet.com/blog/why-i-queue-emails</guid>
      <pubDate>Thu, 15 Jan 2026 00:00:00 GMT</pubDate>
      <description>A 12k-row CSV taught me that sending email synchronously is a request-handling bug in disguise. Here&apos;s the rewrite that fixed it.</description>
      <category>Backend</category>
      <category>Queues</category>
      <category>Redis</category>
    </item>
    <item>
      <title>How I cut a Next.js page&apos;s TBT from 510ms to 300ms in one PR</title>
      <link>https://www.modimeet.com/blog/cutting-tbt-by-41-percent</link>
      <guid isPermaLink="true">https://www.modimeet.com/blog/cutting-tbt-by-41-percent</guid>
      <pubDate>Mon, 10 Nov 2025 00:00:00 GMT</pubDate>
      <description>HacktoberFest 2025. PrivGPT-Studio&apos;s docs page was sluggish on mobile. Profiling found three problems I would never have guessed from reading the code.</description>
      <category>Performance</category>
      <category>Next.js</category>
      <category>Open Source</category>
    </item>
  </channel>
</rss>
