<?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: Abrham Taye</title>
    <description>The latest articles on DEV Community by Abrham Taye (@abrham_taye_b61dddeb30814).</description>
    <link>https://dev.to/abrham_taye_b61dddeb30814</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.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4247383%2F5c4c250b-435b-48ab-b779-6fcabbf9ddd8.png</url>
      <title>DEV Community: Abrham Taye</title>
      <link>https://dev.to/abrham_taye_b61dddeb30814</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/abrham_taye_b61dddeb30814"/>
    <language>en</language>
    <item>
      <title>Building Content Monitoring System</title>
      <dc:creator>Abrham Taye</dc:creator>
      <pubDate>Mon, 16 Nov 2026 10:01:28 +0000</pubDate>
      <link>https://dev.to/abrham_taye_b61dddeb30814/building-content-monitoring-system-1bo4</link>
      <guid>https://dev.to/abrham_taye_b61dddeb30814/building-content-monitoring-system-1bo4</guid>
      <description>&lt;h1&gt;
  
  
  ai #nlp #africa #opensource
&lt;/h1&gt;




&lt;p&gt;I'm building SENAY, a system that monitors public Amharic social media for harmful narratives around GBV, FGM and sexual health misinformation. It will give a civil society organization a weekly summary of what's spreading with a  monitoring options.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Chatbot vs. monitor.&lt;/strong&gt;&lt;br&gt;
My first idea was an Amharic health chatbot: RAG over documents, a Telegram bot, done. But that's mostly wiring up a model someone else trained. I chose the monitor because it means collecting data, defining labels, fine-tuning, and analyzing failures. That's the harder work, and the more useful kind.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Standalone, not integrated.&lt;/strong&gt;&lt;br&gt;
The partner organization already has digital products. Instead of working inside a codebase I don't know, I'm building a separate system that only feeds them information. In a short project, that saves weeks.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Platform access limits the scope.&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Telegram: public channels are easy to read, so I'm starting here.&lt;/li&gt;
&lt;li&gt;Facebook: only public pages are realistic. Private groups are out of reach.&lt;/li&gt;
&lt;li&gt;Twitter/X: API costs make it impractical, so I'm skipping it and will say so in the write-up.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;"Harmful" shouldn't be defined by the developer alone.&lt;/strong&gt; &lt;br&gt;
A post discussing FGM and a post defending it are very different things. The partner's staff know their communities, so we're writing the annotation guide together, and I'll measure how often annotators agree before training anything.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Open questions.&lt;/strong&gt; &lt;br&gt;
I don't know yet whether multilingual models like AfroXLMR will handle slang and code-switching in Amharic, how much labeled data I can realistically get, or whether the weekly digest will actually be used. That last one matters most, since a tool nobody uses is just a demo.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Plan:&lt;/strong&gt; &lt;br&gt;
compare baselines (keyword matching, zero-shot LLM, fine-tuned models), document failure cases, and release the dataset if ethics review allows.&lt;/p&gt;

&lt;p&gt;If you've worked on NLP for Amharic or other low-resource languages, I'd like to hear what challenges you ran into.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>opensource</category>
      <category>socialmedia</category>
    </item>
  </channel>
</rss>
