<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Ruksh E Ibadat — Insights</title>
    <link>https://ruksheibadat.com/insights/</link>
    <description>Long-form engineering writing across agentic AI in production, planetary AI, healthcare AI under regulatory constraint, and SEN education technology.</description>
    <language>en-us</language>
    <copyright>© 2026 Ruksh E Ibadat. All rights reserved.</copyright>
    <managingEditor>ruksh.e.ibadat@icloud.com (Ruksh E Ibadat)</managingEditor>
    <webMaster>ruksh.e.ibadat@icloud.com (Ruksh E Ibadat)</webMaster>
    <lastBuildDate>Thu, 14 May 2026 12:00:00 +0000</lastBuildDate>
    <atom:link href="https://ruksheibadat.com/rss.xml" rel="self" type="application/rss+xml" />

    <item>
      <title>Why MARS-MINDS uses five models and a controller, not an end-to-end network</title>
      <link>https://ruksheibadat.com/insights/why-mars-minds-uses-five-models</link>
      <guid isPermaLink="true">https://ruksheibadat.com/insights/why-mars-minds-uses-five-models</guid>
      <pubDate>Thu, 14 May 2026 12:00:00 +0000</pubDate>
      <author>ruksh.e.ibadat@icloud.com (Ruksh E Ibadat)</author>
      <category>Planetary AI</category>
      <description><![CDATA[The load-bearing architecture decision behind the 99.91 per cent Mission Planning benchmark — why an end-to-end network would have lost three points and broken the upgrade path. 11 min read.]]></description>
    </item>

    <item>
      <title>HIPAA-aligned agentic AI: what the audit log has to know</title>
      <link>https://ruksheibadat.com/insights/hipaa-agentic-audit-log</link>
      <guid isPermaLink="true">https://ruksheibadat.com/insights/hipaa-agentic-audit-log</guid>
      <pubDate>Thu, 30 Apr 2026 12:00:00 +0000</pubDate>
      <author>ruksh.e.ibadat@icloud.com (Ruksh E Ibadat)</author>
      <category>Healthcare AI</category>
      <description><![CDATA[The audit-log architecture that survived clinical governance review for DRIS v1.0. Every tool call, every prompt, every routing decision — SHA-256 signed, per-tenant, replay-able. 9 min read.]]></description>
    </item>

    <item>
      <title>IEP generation across thirty SEN boards: the schema we ended up with</title>
      <link>https://ruksheibadat.com/insights/iep-generation-schema</link>
      <guid isPermaLink="true">https://ruksheibadat.com/insights/iep-generation-schema</guid>
      <pubDate>Thu, 16 Apr 2026 12:00:00 +0000</pubDate>
      <author>ruksh.e.ibadat@icloud.com (Ruksh E Ibadat)</author>
      <category>SEN Education</category>
      <description><![CDATA[The board-aligned JSON schema and citation layer behind LUMINA's sub-three-second IEP generation. Why we rejected a single canonical schema and built a per-board translation layer instead. 10 min read.]]></description>
    </item>
  </channel>
</rss>
