<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Writing by Sergio A. Mora Pardo</title><link>https://sergiomorapardo.github.io/writing/</link><description>Technical notes on applied machine learning and AI systems.</description><generator>Hugo</generator><language>en-US</language><atom:link href="https://sergiomorapardo.github.io/writing/index.xml" rel="self" type="application/rss+xml"/><item><title>Cross-validation that matches the data</title><link>https://sergiomorapardo.github.io/writing/cross-validation-that-matches-the-data/</link><guid>https://sergiomorapardo.github.io/writing/cross-validation-that-matches-the-data/</guid><pubDate>Sun, 06 Sep 2026 00:00:00 -0500</pubDate><description>A defensible model-selection workflow: choose splits from the data-generating process, keep preprocessing inside each fold, and separate tuning from evaluation.</description></item><item><title>From co-likes to candidates with Node2Vec</title><link>https://sergiomorapardo.github.io/writing/node2vec-recommendations/</link><guid>https://sergiomorapardo.github.io/writing/node2vec-recommendations/</guid><pubDate>Sun, 06 Sep 2026 00:00:00 -0500</pubDate><description>A practical route from MovieLens interactions to graph embeddings, along with the evidence and limits needed before calling the result a recommender.</description></item></channel></rss>