<?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/956a1843c6f0445490b97879d4f330f8&quot; frameborder=&quot;0&quot; width=&quot;3456&quot; height=&quot;2592&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>2592</height><width>3456</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>2592</thumbnail_height><thumbnail_width>3456</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/956a1843c6f0445490b97879d4f330f8-189516841116ad6a.gif</thumbnail_url><duration>240.737</duration><title>Meet W&amp;amp;B Agent: An AI Research Assistant for Your ML Experiments</title><description>Hi, I am Julia from Weights and Bices, and today I introduced the Weights and Bices Agement, an AI research assistant built into your workspace with access to your experiments and context. It can summarize projects, surface outlier runs, analyze patterns, and even generate visualizations and a clickable report with recommended next experiments. In my demo, I fine tuned an open source email agent using the quad model, and it identified both positive and negative outlier panels. I also asked it to recommend the next job and explained how you can have it launch jobs overnight. If you want access, please contact me over email for this private preview.</description></oembed>