<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Architecture on René Zander | AI Automation Consultant</title><link>https://renezander.com/tags/architecture/</link><description>Recent content in Architecture on René Zander | AI Automation Consultant</description><generator>Hugo</generator><language>en</language><lastBuildDate>Fri, 10 Apr 2026 08:00:00 +0200</lastBuildDate><atom:link href="https://renezander.com/tags/architecture/index.xml" rel="self" type="application/rss+xml"/><item><title>Production AI Agent Architecture: Patterns That Actually Ship</title><link>https://renezander.com/guides/production-ai-agent-architecture/</link><pubDate>Fri, 10 Apr 2026 08:00:00 +0200</pubDate><guid>https://renezander.com/guides/production-ai-agent-architecture/</guid><description>&lt;p>Most agent tutorials end at &amp;ldquo;the model calls tools in a loop, done.&amp;rdquo; That works for a demo. It falls apart the first time a tool 500s, a user asks something off-script, or the token bill crosses $20 on a single task. Production AI agent architecture is the set of patterns that keep that loop alive when reality hits.&lt;/p>
&lt;p>I run 10 agents in production right now. Bash scripts calling &lt;code>claude -p&lt;/code>, scheduled via systemd timers, reporting outcomes to Telegram. Not fancy. They ship work every day because the architecture around the loop is boring and deliberate. This guide is that playbook: the patterns, the must-haves, the anti-patterns, and the opinionated verdicts on what to use when.&lt;/p></description></item></channel></rss>