<?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/"
     xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>Syncorix Blog — Notes on African Healthcare AI</title>
    <link>https://www.syncorixglobal.com/blog</link>
    <atom:link href="https://www.syncorixglobal.com/rss.xml" rel="self" type="application/rss+xml"/>
    <description>Perspectives on clinical data, model training, and the health systems from the team building AI healthcare infrastructure for Africa.</description>
    <language>en-GB</language>
    <copyright>© 2026 Syncorix Global. All rights reserved.</copyright>
    <managingEditor>info@syncorixglobal.com (Syncorix Editorial)</managingEditor>
    <webMaster>info@syncorixglobal.com (Syncorix Editorial)</webMaster>
    <lastBuildDate>Sun, 19 Apr 2026 00:00:00 +0000</lastBuildDate>
    <pubDate>Sun, 19 Apr 2026 00:00:00 +0000</pubDate>
    <ttl>1440</ttl>
    <image>
      <url>https://www.syncorixglobal.com/assets/icons/icon-512.png</url>
      <title>Syncorix</title>
      <link>https://www.syncorixglobal.com/blog</link>
      <width>512</width>
      <height>512</height>
    </image>

    <item>
      <title>Building a voice consultation engine pharmacists trust</title>
      <link>https://www.syncorixglobal.com/blog/voice-consultation-engine</link>
      <guid isPermaLink="true">https://www.syncorixglobal.com/blog/voice-consultation-engine</guid>
      <pubDate>Sun, 19 Apr 2026 00:00:00 +0000</pubDate>
      <dc:creator>Syncorix Editorial</dc:creator>
      <category>Engineering</category>
      <description><![CDATA[The audio pipeline — RNNoise → Silero VAD → smart-turn-v3 → NLI matcher — that replaced a typed pharmacy consultation form with voice, and the bug that taught us a length-ratio guard.]]></description>
      <content:encoded><![CDATA[
        <p>Three problems, in order: hear it, know when the speaker stopped, know what they meant.</p>
        <p>Every new model ships in shadow first — parallel, telemetry-only, no influence on user-visible behaviour. The matcher we use is the one we earned the right to use.</p>
        <p><a href="https://www.syncorixglobal.com/blog/voice-consultation-engine">Read the full piece →</a></p>
      ]]></content:encoded>
    </item>

    <item>
      <title>Before AI Could Matter in African Healthcare</title>
      <link>https://www.syncorixglobal.com/blog/before-ai-could-matter-in-african-healthcare</link>
      <guid isPermaLink="true">https://www.syncorixglobal.com/blog/before-ai-could-matter-in-african-healthcare</guid>
      <pubDate>Sun, 19 Apr 2026 00:00:00 +0000</pubDate>
      <dc:creator>Syncorix Editorial</dc:creator>
      <category>Perspectives</category>
      <description><![CDATA[Before clinical intelligence can become useful, the harder problem is building the data foundations that healthcare systems across Africa still lack.]]></description>
      <content:encoded><![CDATA[
        <p>The conversation about clinical AI in African healthcare moves too quickly to the question of which model to deploy. But a model trained on fragmented, non-longitudinal data will produce confident noise, not reliable intelligence.</p>
        <p>Rwanda's National Health Intelligence Centre, launched in April 2025, now predicts disease outbreaks three months in advance. That capability did not arrive with the technology. It followed years of deliberate investment in connected health data infrastructure.</p>
        <p><a href="https://www.syncorixglobal.com/blog/before-ai-could-matter-in-african-healthcare">Read the full piece →</a></p>
      ]]></content:encoded>
    </item>

  </channel>
</rss>
