<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Qwen on Johnson Lai · Forward Deployed Engineer Malaysia/Singapore</title><link>http://superoo7.com/tags/qwen/</link><description>Recent content in Qwen on Johnson Lai · Forward Deployed Engineer Malaysia/Singapore</description><generator>Hugo</generator><language>en</language><lastBuildDate>Fri, 22 May 2026 20:00:00 +0800</lastBuildDate><atom:link href="http://superoo7.com/tags/qwen/index.xml" rel="self" type="application/rss+xml"/><item><title>My Multica workflow: a human + local-LLM agent fleet</title><link>http://superoo7.com/posts/multica-local-llm-fde-workflow/</link><pubDate>Fri, 22 May 2026 20:00:00 +0800</pubDate><guid>http://superoo7.com/posts/multica-local-llm-fde-workflow/</guid><description>&lt;h2 id="why-agent-first"&gt;
 Why agent-first
 &lt;a class="heading-link" href="#why-agent-first"&gt;
 &lt;i class="fa-solid fa-link" aria-hidden="true" title="Link to heading"&gt;&lt;/i&gt;
 &lt;span class="sr-only"&gt;Link to heading&lt;/span&gt;
 &lt;/a&gt;
&lt;/h2&gt;
&lt;p&gt;Most teams still treat AI agents like fancy autocomplete: open IDE, ask Copilot, paste, repeat. That&amp;rsquo;s a tool, not a workflow. We saw the hype around tools like &lt;a href="https://openclaw.ai/" class="external-link" target="_blank" rel="noopener"&gt;OpenClaw&lt;/a&gt; and &lt;a href="https://hermes-agent.nousresearch.com/" class="external-link" target="_blank" rel="noopener"&gt;Hermes Agent&lt;/a&gt;, but many users install them, poke around, and then get lost because they do not have an operating loop for using agents effectively.&lt;/p&gt;
&lt;p&gt;As a &lt;a href="http://superoo7.com/forward-deployed-engineer/" &gt;Forward Deployed Engineer&lt;/a&gt;, my job is to redesign the workflow itself: agents handle the repeatable execution, humans review the highest-leverage decisions, and the system keeps shipping.&lt;/p&gt;</description></item></channel></rss>