<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>roughcuts.dev</title><description>A data scientist&apos;s blog on AI, agentic workflows, and the messy reality of building with LLMs.</description><link>https://roughcuts.dev/</link><item><title>Agentic AI workflows for Data Science - Types, risks and mitigation</title><link>https://roughcuts.dev/blog/current-state-of-agentic-workflows-for-data-science/</link><guid isPermaLink="true">https://roughcuts.dev/blog/current-state-of-agentic-workflows-for-data-science/</guid><description>A data scientist&apos;s musings on the fragility of black-box agentic systems</description><pubDate>Sat, 09 May 2026 00:00:00 GMT</pubDate></item><item><title>Useful, Until It Isn&apos;t: LLMs for Data Science</title><link>https://roughcuts.dev/blog/from-zero-to-one/</link><guid isPermaLink="true">https://roughcuts.dev/blog/from-zero-to-one/</guid><description>A hands-on data scientist&apos;s honest take on what actually works when you put AI agents to the test on real workflows — and what still needs a human in the loop.</description><pubDate>Tue, 21 Apr 2026 00:00:00 GMT</pubDate></item></channel></rss>