<?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 Devesh Nath Jha</title><link>https://deveshnathjha.github.io/posts/</link><description>Recent content in Posts on Devesh Nath Jha</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Tue, 28 Apr 2026 14:30:00 +0530</lastBuildDate><atom:link href="https://deveshnathjha.github.io/posts/index.xml" rel="self" type="application/rss+xml"/><item><title>Why Naive RAG Fails in Healthcare and What Matters in Clinical RAG</title><link>https://deveshnathjha.github.io/posts/why-naive-rag-fails-healthcare/</link><pubDate>Tue, 28 Apr 2026 14:30:00 +0530</pubDate><guid>https://deveshnathjha.github.io/posts/why-naive-rag-fails-healthcare/</guid><description>An in-depth guide on why naive Retrieval-Augmented Generation (RAG) fails in clinical healthcare systems, and the production-grade design patterns (parent-child chunking, cross-encoder reranking, and dynamic LLMOps observability) needed to fix them.</description></item></channel></rss>