<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>thoughts on</title><link>https://cdong.us/categories/thoughts/</link><description>Recent content in thoughts on</description><generator>Hugo -- gohugo.io</generator><lastBuildDate>Fri, 03 Jul 2026 09:19:42 +0100</lastBuildDate><atom:link href="https://cdong.us/categories/thoughts/index.xml" rel="self" type="application/rss+xml"/><item><title>My 5-9 Technical Thoughts</title><link>https://cdong.us/blog/my-5-9-technical-thoughts/</link><pubDate>Fri, 03 Jul 2026 09:19:42 +0100</pubDate><guid>https://cdong.us/blog/my-5-9-technical-thoughts/</guid><description>One thing that LLMs — or any technology, really — still haven&amp;rsquo;t completely solved is processing PDFs.
Tax returns, medical records, simple ACORD insurance forms: they&amp;rsquo;re all potential gold mines if you can reliably pull the information out of them. My side venture tackled exactly this by processing medical records. You use a mix of prompts with Pydantic to structure the data and capture the fields you care about, then run a RAG approach over several key topics.</description></item></channel></rss>