<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Miha Peroša</title><description>Notes on AI engineering, LLM reliability, and backend systems.</description><link>https://mihaperosa.com/</link><item><title>AWS Blocks and the hard 20%</title><link>https://mihaperosa.com/writing/aws-blocks-leaky-abstraction/</link><guid isPermaLink="true">https://mihaperosa.com/writing/aws-blocks-leaky-abstraction/</guid><description>A framework that makes local RAG feel like magic, until you deploy. Where the abstraction held, where it leaked, and what I&apos;d reach for it for.</description><pubDate>Sat, 27 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Making an AI agent reliable</title><link>https://mihaperosa.com/writing/making-an-ai-agent-reliable/</link><guid isPermaLink="true">https://mihaperosa.com/writing/making-an-ai-agent-reliable/</guid><description>Three times building one agent, the same bug bit me. Each time the fix was the same principle: never let the model hold a value that has to be exact.</description><pubDate>Mon, 22 Jun 2026 00:00:00 GMT</pubDate></item></channel></rss>