<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Adalat AI on Kavya Manohar</title>
    <link>https://kavyamanohar.com/tags/adalat-ai/</link>
    <description>Recent content in Adalat AI on Kavya Manohar</description>
    <generator>Hugo</generator>
    <language>en</language>
    <managingEditor>sakhi.kavya@gmail.com (Kavya Manohar)</managingEditor>
    <webMaster>sakhi.kavya@gmail.com (Kavya Manohar)</webMaster>
    <lastBuildDate>Sat, 28 Mar 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://kavyamanohar.com/tags/adalat-ai/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>One Year In</title>
      <link>https://kavyamanohar.com/post/work-anniversary/</link>
      <pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><author>sakhi.kavya@gmail.com (Kavya Manohar)</author>
      <guid>https://kavyamanohar.com/post/work-anniversary/</guid>
      <description>March. One year since I joined Adalat AI.&#xA;I have already written about the work itself — the courtrooms, the languages, the gap between benchmarks and real benches. This post is not about that. This is about what happened to life around the job.&#xA;Off the Clock I came in carrying a particular kind of guilt — the kind that early career researchers accumulate quietly over years. The feeling of always being one paper short, one grant proposal behind, perpetually under-delivering against some invisible standard.</description>
    </item>
    <item>
      <title>From Benchmarks to Benches</title>
      <link>https://kavyamanohar.com/post/adalat-work/</link>
      <pubDate>Wed, 29 Oct 2025 00:00:00 +0000</pubDate><author>sakhi.kavya@gmail.com (Kavya Manohar)</author>
      <guid>https://kavyamanohar.com/post/adalat-work/</guid>
      <description>I haven&amp;rsquo;t posted much about my job at Adalat AI. The team I&amp;rsquo;m part of builds dictation systems for the Indian court rooms, and initially, I thought it would be a straightforward extension of my PhD research. I was wrong.&#xA;Unlike academic work where success means beating SOTA on benchmark datasets, the reality at Adalat AI demanded something more: a deep understanding of India&amp;rsquo;s diverse courtroom dynamics, legal workflows, and the intricate linguistic landscape where regional languages and English constantly intertwine.</description>
    </item>
  </channel>
</rss>
