<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Posts on Boul'Og</title><link>https://darylbdr.github.io/posts/</link><description>Recent content in Posts on Boul'Og</description><generator>Hugo</generator><language>en-us</language><atom:link href="https://darylbdr.github.io/posts/index.xml" rel="self" type="application/rss+xml"/><item><title>MLP from scratch (WIP)</title><link>https://darylbdr.github.io/posts/mlp_from_scratch/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://darylbdr.github.io/posts/mlp_from_scratch/</guid><description>&lt;p&gt;Here I implemented a MLP from scratch where &lt;strong&gt;The goal was to be able to generate fictional character names&lt;/strong&gt;. The full code is available in this &lt;a href="https://github.com/darylbdr/AI-Sandbox/blob/main/MLP/MLP_from_scratch.ipynb"&gt;github repo&lt;/a&gt;.&lt;/p&gt;
&lt;p&gt;I followed the approach introduced in the &lt;a href="https://www.jmlr.org/papers/volume3/bengio03a/bengio03a.pdf"&gt;Bengio et al. 2003 MLP language model paper&lt;/a&gt; where the MLP architecture proposed in this paper is the following :
&lt;figure&gt;
&lt;img loading="lazy" src="https://darylbdr.github.io/images/architecture.png"
alt="Figure 1: the MLP architecture according to the paper"/&gt; &lt;figcaption&gt;
&lt;p&gt;Figure 1: the MLP architecture according to the paper&lt;/p&gt;</description></item><item><title>Local GitLab AI Coding Agent</title><link>https://darylbdr.github.io/posts/gitlab_coding_agent/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://darylbdr.github.io/posts/gitlab_coding_agent/</guid><description>&lt;p&gt;I implemented an autonomous AI coding agent designed to resolve GitLab issues using local LLMs via &lt;a href="https://ollama.com/"&gt;Ollama&lt;/a&gt; and the &lt;a href="https://python-gitlab.readthedocs.io/en/stable/"&gt;python-gitlab&lt;/a&gt; API.&lt;/p&gt;
&lt;p&gt;By leveraging native &lt;strong&gt;Function Calling&lt;/strong&gt;, the agent independently explores repository structures, reads files, implements fixes on dedicated branches, and submits Merge Requests—keeping your codebase 100% private and on-premise.&lt;/p&gt;
&lt;h2 id="features"&gt;Features&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Monitors GitLab issues labeled with &lt;code&gt;ai:agent&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Analyzes the issue title and description.&lt;/li&gt;
&lt;li&gt;Interacts with the repository (reads files, lists directories).&lt;/li&gt;
&lt;li&gt;Creates a dedicated branch for the fix.&lt;/li&gt;
&lt;li&gt;Generates code fixes and commits them.&lt;/li&gt;
&lt;li&gt;Opens a Merge Request when the fix is ready.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;&lt;em&gt;The full code is available on this &lt;a href="https://github.com/darylbdr/Local-Gitlab-Coding-Agent"&gt;github repo&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;</description></item></channel></rss>