
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>Navigating the Digital Landscape</title>
      <link>https://mibrahim210.vercel.app/blog</link>
      <description>Tech, Thoughts, and Tales</description>
      <language>en-us</language>
      <managingEditor>mibrahim@rational-analytiics.com (Md Ibrahim)</managingEditor>
      <webMaster>mibrahim@rational-analytiics.com (Md Ibrahim)</webMaster>
      <lastBuildDate>Mon, 28 Sep 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://mibrahim210.vercel.app/tags/mlops/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://mibrahim210.vercel.app/blog/artificial-intelligence/medical-imaging/invariant-ibm-bob-hackathon-september</guid>
    <title>Invariant: When the Tests Pass but the ML Evaluation Is Wrong</title>
    <link>https://mibrahim210.vercel.app/blog/artificial-intelligence/medical-imaging/invariant-ibm-bob-hackathon-september</link>
    <description>How I built Invariant with IBM Bob to catch patient-level data leakage during code review, verify a fix, and prevent the bug from returning.</description>
    <pubDate>Mon, 28 Sep 2026 00:00:00 GMT</pubDate>
    <author>mibrahim@rational-analytiics.com (Md Ibrahim)</author>
    <category>machine-learning</category><category>medical-imaging</category><category>data-leakage</category><category>mlops</category><category>code-review</category><category>ibm-bob</category>
  </item>

    </channel>
  </rss>
