<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Jasper Lu</title><description>Designing software in the agentic era.</description><link>https://jasperlu.com/</link><item><title>Training Search Agents with GRPO</title><link>https://jasperlu.com/blog/training-search-agents-grpo/</link><guid isPermaLink="true">https://jasperlu.com/blog/training-search-agents-grpo/</guid><description>Reinforcement learning for search over a financial-filings corpus: the dataset, the runs, and what the reward curves actually say.</description><pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate><category>rl</category><category>search</category><category>grpo</category></item></channel></rss>