<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Evgeny Osipchuk</title><description>Lead AI engineer writing about agent architectures, LLM safety, and the failure modes that live above the model.</description><link>https://evgenyosipchuk.com/</link><language>en-us</language><item><title>AI Mentor for Sales Managers: From Daily Reports to an Interactive Simulator</title><link>https://evgenyosipchuk.com/writing/2026-05-ai-mentor-for-sales-managers/</link><guid isPermaLink="true">https://evgenyosipchuk.com/writing/2026-05-ai-mentor-for-sales-managers/</guid><description>How a workhorse LLM tool went from daily email reports to an in-app simulator: prompt selection, a failed A/B test, prompt chaining, and what I took away from it.</description><pubDate>Wed, 20 May 2026 00:00:00 GMT</pubDate><category>llm</category><category>ai-product</category><category>speech-analytics</category><category>prompt-engineering</category><category>ml-pilot</category><category>feedback-loop</category></item></channel></rss>