https://manuel-delverme.github.io/Recent content on Quantified MindHugo -- 0.147.0en-usSat, 10 Jan 2026 00:00:00 +0000https://manuel-delverme.github.io/about/Sat, 10 Jan 2026 00:00:00 +0000https://manuel-delverme.github.io/about/<p>PhD student under Pierre-Luc Bacon and Doina Precup. I split my time between research and building.</p> <h2 id="research">Research</h2> <p>Working on questions like:</p> <ul> <li>Do we really need silicon-based computation?</li> <li>Can we build interfaces that adapt to users rather than the other way around?</li> <li>Can we train bigger neural networks in a distributed fashion?</li> </ul> <p>Tools like <a href="https://github.com/manuel-delverme/experiment_buddy">experiment_buddy</a> aim to make AI researchers&rsquo; workflows smoother.</p> <p>Book a chat: <a href="https://calendly.com/manuel-delverme/15min">calendly.com/manuel-delverme/15min</a></p> <hr> <h2 id="quantified-self">Quantified Self</h2> <p>I track everything. 26 data sources spanning 13+ years: Spotify plays, GitHub commits, browser history, location data, messages, calendar events, finance transactions, health metrics.</p>https://manuel-delverme.github.io/posts/skip-rate-signal/Sat, 10 Jan 2026 00:00:00 +0000https://manuel-delverme.github.io/posts/skip-rate-signal/I analyzed 96,000 Spotify plays against 6,000 GitHub commits. Skip rate above 20% predicts a 55% drop in next-day output.https://manuel-delverme.github.io/posts/constrained-neural-networks/Thu, 14 Jan 2021 00:00:00 +0000https://manuel-delverme.github.io/posts/constrained-neural-networks/A survey of approaches to divide the parameter estimation problem for neural networks into simpler sub-problems using constrained optimization.