<?xml version="1.0" encoding="utf-8"?>
<?xml-stylesheet href="/rss/style.xsl" media="screen" type="text/xsl"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Sushan Poudel | Blog</title><link>https://sushanpoudel.com/blog/</link><description>Articles on AI engineering and Python backend development: RAG pipelines, LLM agents, semantic caching, Django and FastAPI APIs and system design, with code.</description><atom:link href="https://sushanpoudel.com/rss/" rel="self"/><language>en-us</language><lastBuildDate>Wed, 08 Apr 2026 14:53:25 +0000</lastBuildDate><item><title>Building Your First AI Agent (That Actually Does Something Useful)</title><link>https://sushanpoudel.com/blog/building-your-first-ai-agent/</link><description>Build a working AI agent with LangGraph: two tools, a clear prompt, a step limit and a loop that knows when to stop.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Sushan Poudel</dc:creator><pubDate>Wed, 08 Apr 2026 14:53:25 +0000</pubDate><guid>https://sushanpoudel.com/blog/building-your-first-ai-agent/</guid><category>AI agents</category><category>LLM</category><category>LangGraph</category><category>Python</category></item><item><title>Data Leakage Begins in Preprocessing (And Most People Miss It)</title><link>https://sushanpoudel.com/blog/data-leakage-in-preprocessing/</link><description>Four preprocessing mistakes that leak test data into training, why validation hides them, and how to split first and use pipelines.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Sushan Poudel</dc:creator><pubDate>Mon, 16 Feb 2026 15:28:54 +0000</pubDate><guid>https://sushanpoudel.com/blog/data-leakage-in-preprocessing/</guid><category>Data science</category><category>Machine learning</category><category>Python</category><category>scikit-learn</category></item></channel></rss>