<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Kamil Tustanowski</title>
    <description>The latest articles on DEV Community by Kamil Tustanowski (@kamil_tustanowski_8d33739).</description>
    <link>https://dev.to/kamil_tustanowski_8d33739</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3766744%2Fa32bf4cc-bf79-48c2-8c20-9b0f1884516c.jpg</url>
      <title>DEV Community: Kamil Tustanowski</title>
      <link>https://dev.to/kamil_tustanowski_8d33739</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/kamil_tustanowski_8d33739"/>
    <language>en</language>
    <item>
      <title>AI Agents 101: From Concept to Code (No Frameworks Required)</title>
      <dc:creator>Kamil Tustanowski</dc:creator>
      <pubDate>Wed, 11 Feb 2026 16:16:47 +0000</pubDate>
      <link>https://dev.to/kamil_tustanowski_8d33739/ai-agents-101-from-concept-to-code-no-frameworks-required-5cgd</link>
      <guid>https://dev.to/kamil_tustanowski_8d33739/ai-agents-101-from-concept-to-code-no-frameworks-required-5cgd</guid>
      <description>&lt;p&gt;Most tutorials about AI agents rely heavily on frameworks.&lt;br&gt;
LangChain.&lt;br&gt;
AutoGen.&lt;br&gt;
CrewAI.&lt;/p&gt;

&lt;p&gt;They’re powerful — but they abstract away the most important part:&lt;br&gt;
How an AI agent actually works under the hood.&lt;/p&gt;

&lt;p&gt;So I decided to build one from scratch.&lt;/p&gt;

&lt;p&gt;No frameworks.&lt;br&gt;
Just Python, an LLM, and a clear decision loop.&lt;/p&gt;

&lt;p&gt;What This Article Covers&lt;br&gt;
In this write-up, I walk through:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;What makes an AI agent different from a plain LLM&lt;/li&gt;
&lt;li&gt;The core “Think → Act → Observe” loop&lt;/li&gt;
&lt;li&gt;How tool calling works conceptually&lt;/li&gt;
&lt;li&gt;How to structure a simple agent controller&lt;/li&gt;
&lt;li&gt;How to let an agent browse and retrieve information&lt;/li&gt;
&lt;li&gt;A fully working example at the end&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The goal wasn’t to build something production-ready.&lt;/p&gt;

&lt;p&gt;The goal was to understand the mechanics deeply.&lt;/p&gt;

&lt;p&gt;Why Build Without Frameworks?&lt;br&gt;
Frameworks are great for speed.&lt;br&gt;
But if you:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Want to customize behavior&lt;/li&gt;
&lt;li&gt;Need fine-grained control&lt;/li&gt;
&lt;li&gt;Are debugging weird agent decisions&lt;/li&gt;
&lt;li&gt;Or simply want to understand what's happening&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Then building one manually is incredibly valuable.&lt;br&gt;
Once you understand the loop, frameworks make much more sense.&lt;/p&gt;

&lt;p&gt;If you're interested in how AI agents really work and want a practical, code-first explanation, you can read the full article here:&lt;/p&gt;

&lt;p&gt;👉 &lt;a href="https://medium.com/@kamil.tustanowski/ai-agents-101-from-concept-to-code-no-frameworks-required-2dfdaf66b6c1" rel="noopener noreferrer"&gt;https://medium.com/@kamil.tustanowski/ai-agents-101-from-concept-to-code-no-frameworks-required-2dfdaf66b6c1&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Would love feedback — especially from people building agent systems in real-world projects.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>python</category>
      <category>opensource</category>
      <category>machinelearning</category>
    </item>
  </channel>
</rss>
