<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
    <channel>
        <title>Peter McCann Strain — Build Notes</title>
        <link>https://petermccannstrain.com/blog</link>
        <description>Thoughts on AI engineering, multi-agent systems, computational biology, and production ML.</description>
        <lastBuildDate>Sat, 20 Jun 2026 06:39:03 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <language>en-GB</language>
        <copyright>© 2026 Peter McCann Strain</copyright>
        <item>
            <title><![CDATA[Bringing Old Family Photos To Life With AI]]></title>
            <link>https://petermccannstrain.com/blog/bringing-old-family-photos-to-life</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/bringing-old-family-photos-to-life</guid>
            <pubDate>Thu, 21 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[A technical log of an AI-assisted family film: colour-managed restoration, image-to-video animation, identity preservation, and the local pipeline that got built and then bypassed.]]></description>
            <author>Peter McCann Strain</author>
            <category>Computer Vision</category>
            <category>Production AI</category>
        </item>
        <item>
            <title><![CDATA[From Whole-Slide Images To Biological Structure: My Oxford DPhil In Medical AI]]></title>
            <link>https://petermccannstrain.com/blog/medical-ai-oxford-dphil</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/medical-ai-oxford-dphil</guid>
            <pubDate>Wed, 20 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[A technical account of my Oxford DPhil: 3 million unlabelled cell nuclei from mouse placental histology, an 89% supervised classifier, near-perfect clustering of 8 cell types, and a self-supervised experiment that fell short of its goal.]]></description>
            <author>Peter McCann Strain</author>
            <category>Medical AI</category>
            <category>Computer Vision</category>
            <category>Research Methods</category>
        </item>
        <item>
            <title><![CDATA[Equal Pay as a Data Operation: A Union's Claims Base and a Valuation Engine It Can Argue With]]></title>
            <link>https://petermccannstrain.com/blog/settlemise-equal-pay-valuations</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/settlemise-equal-pay-valuations</guid>
            <pubDate>Wed, 20 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[How Datamise built the equal-pay work behind GMB Union's campaigns — an owned, relational claims base and the 38-finding audit that hardened it, plus SettleMise, a valuation engine where every number is traceable and a reviewer can take it apart.]]></description>
            <author>Peter McCann Strain</author>
            <category>Founder Engineering</category>
            <category>Legal Infrastructure</category>
        </item>
        <item>
            <title><![CDATA[Removing the LLM From the Hot Path: Building SThree’s Skills Taxonomy Pipeline]]></title>
            <link>https://petermccannstrain.com/blog/removing-the-llm-from-skills-taxonomy</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/removing-the-llm-from-skills-taxonomy</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[A 30-experiment programme that replaced a per-skill LLM validator with a zero-LLM retrieval-and-filter cascade and a 7 kB classifier head — 180× cheaper, ~100× faster, and within ~2pp of the model it replaced.]]></description>
            <author>Peter McCann Strain</author>
            <category>Production AI</category>
            <category>LLM Cost Optimization</category>
            <category>AI Evaluation</category>
            <category>Recruitment AI</category>
        </item>
        <item>
            <title><![CDATA[The Retrieval Bottleneck: A Controlled Comparison of 13 Deep-Research Architectures]]></title>
            <link>https://petermccannstrain.com/blog/retrieval-bottleneck-deep-research</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/retrieval-bottleneck-deep-research</guid>
            <pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[A controlled benchmark of 13 deep-research agent architectures on one shared tool layer — what happens to orchestration claims when you fix the model, the tools, and the judging protocol, and read the statistics honestly.]]></description>
            <author>Peter McCann Strain</author>
            <category>AI Evaluation</category>
            <category>Research Methods</category>
            <category>Multi-Agent Systems</category>
        </item>
        <item>
            <title><![CDATA[Satellite Pose Estimation and the Sim-to-Real Domain Gap]]></title>
            <link>https://petermccannstrain.com/blog/deep-learning-space-debris</link>
            <guid isPermaLink="false">https://petermccannstrain.com/blog/deep-learning-space-debris</guid>
            <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[How I approached spacecraft pose estimation at Astroscale: measuring the sim-to-real gap before modelling it, a preprocessing result that flipped sign with placement, the SPS-B architecture proposal, and AstroGAN — an image-translation model that closed the appearance gap while leaving the harder half open.]]></description>
            <author>Peter McCann Strain</author>
            <category>Computer Vision</category>
            <category>Research Methods</category>
        </item>
    </channel>
</rss>