<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/d72ef56097194dd08fdaa734ff64d90f&quot; frameborder=&quot;0&quot; width=&quot;1920&quot; height=&quot;1440&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1440</height><width>1920</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1440</thumbnail_height><thumbnail_width>1920</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/d72ef56097194dd08fdaa734ff64d90f-c271df79b0699678.gif</thumbnail_url><duration>176.753</duration><title>Gradio Small Build Hackathon 2026</title><description>This Loom demonstrates Clean IQ, an on-device pharmacist meant to catch medication and allergy conflicts across multiple patient documents. Kartek explains that Clean IQ is a 3 billion parameter model running on Qwen 2.5 3V on a 3V setup, built to solve the small clinic issue where nurses upload visit summaries and no one checks conflicts document by document. He shows the UI loading three documents, after which the system scans all files, flags a dangerous conflict where the patient is prescribed Aspirin 81 MG but has an aspirin allergy. He also demonstrates asking questions like listing all medications and “what alternative was recommended,” with the agent retrieving and tracing relevant chunks across documents.</description></oembed>