<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Posts on Kaiwen Che</title><link>https://kaiwenche.github.io/posts/</link><description>Recent content in Posts on Kaiwen Che</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Tue, 07 Apr 2026 00:00:00 -0700</lastBuildDate><atom:link href="https://kaiwenche.github.io/posts/index.xml" rel="self" type="application/rss+xml"/><item><title>Can Coding Agents Learn Editorial Taste with RLVR?</title><link>https://kaiwenche.github.io/posts/rlvr-resume-agent/</link><pubDate>Tue, 07 Apr 2026 00:00:00 -0700</pubDate><guid>https://kaiwenche.github.io/posts/rlvr-resume-agent/</guid><description>Training a Qwen3-14B resume revision agent with RLVR and an LLM judge. Tool mastery comes fast, content quality learns slow, and the rubric tensions reveal what reward design really means for creative domains.</description></item></channel></rss>