<?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: Aditya Singh Yadav</title>
    <description>The latest articles on DEV Community by Aditya Singh Yadav (@aadiitya).</description>
    <link>https://dev.to/aadiitya</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%2F4154106%2F9091977c-cb3c-4d95-8e0e-2f3d658a3385.png</url>
      <title>DEV Community: Aditya Singh Yadav</title>
      <link>https://dev.to/aadiitya</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/aadiitya"/>
    <language>en</language>
    <item>
      <title>Pakka? - Plans, with receipts</title>
      <dc:creator>Aditya Singh Yadav</dc:creator>
      <pubDate>Sun, 04 Oct 2026 21:19:52 +0000</pubDate>
      <link>https://dev.to/aadiitya/pakka-plans-with-receipts-2bcj</link>
      <guid>https://dev.to/aadiitya/pakka-plans-with-receipts-2bcj</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-weekend-2026-10-01"&gt;Hacktoberfest Weekend Challenge: Build for a Friend&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;A friend recognized the problem behind Pakka?: important plans get buried in busy group chats, tentative suggestions can look final, and later changes are easy to miss. I built a prototype to turn pasted chat text into reviewable plan candidates.&lt;/p&gt;

&lt;p&gt;I shared Pakka? with my friend, who already had Ollama installed. He ran it locally on his device, got it working, and genuinely liked the project. That was encouraging: it showed me this idea mattered to someone beyond me.&lt;/p&gt;

&lt;p&gt;Pakka? separates confirmed, proposed, unclear, cancelled, and superseded plans. It shows the source lines behind each interpretation, keeps tasks and RSVPs out of calendar events, and requires a person to review an event before export.&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://drive.google.com/drive/folders/1vZmR8WIu4WR2xkfzcchvyYpjZRZQePoA?usp=sharing" rel="noopener noreferrer"&gt;Drive folder&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://github.com/aditya-mnnit/pakka-hf26" rel="noopener noreferrer"&gt;View Pakka? on GitHub&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;Pakka? is a Node.js app with a Gemma 4 integration through Ollama. It asks the model for plan candidates with source quotations and line numbers, then checks that cited text appears in the supplied conversation. A person can edit and review candidates before exporting valid, confirmed calendar events. My friend ran it locally using his Ollama setup, and it worked on his device.&lt;/p&gt;

&lt;p&gt;The automated test suite passes 13 tests covering evidence checks, dates, cancellations, review requirements, and calendar export. This is a small software test suite, not a broad measure of model accuracy.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Ollama let my friend run Pakka? on his own device instead of having to use a hosted model service. In local mode, chat text goes to the local Ollama instance; the optional hosted mode sends it to Google's Gemini API. That gives people a choice about where inference happens. The local run worked for my friend, though I have not measured model accuracy or compared it with a closed model.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>weekendchallenge</category>
      <category>hf26challenge</category>
    </item>
  </channel>
</rss>
