<?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: Debjyoti Saha</title>
    <description>The latest articles on DEV Community by Debjyoti Saha (@debjyoti_saha_63b315289e6).</description>
    <link>https://dev.to/debjyoti_saha_63b315289e6</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%2F4103156%2Ffd2604bb-cb21-47de-9921-d5869cfa1125.png</url>
      <title>DEV Community: Debjyoti Saha</title>
      <link>https://dev.to/debjyoti_saha_63b315289e6</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/debjyoti_saha_63b315289e6"/>
    <language>en</language>
    <item>
      <title>week 4</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Tue, 27 Oct 2026 16:46:18 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/week-4-4bob</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/week-4-4bob</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week4-2026-10-26"&gt;Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

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

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

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

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

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

&lt;h2&gt;
  
  
  My Agent Session
&lt;/h2&gt;

&lt;h2&gt;
  
  
  The Spirit of It
&lt;/h2&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>hi</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Wed, 21 Oct 2026 06:59:26 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/hi-294e</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/hi-294e</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week3-2026-10-19"&gt;Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

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

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

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

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

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

&lt;h2&gt;
  
  
  My Agent Session
&lt;/h2&gt;

&lt;h2&gt;
  
  
  The Lazy Payoff
&lt;/h2&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>CSEHub: Turning Computer Science Articles into an AI-Powered Learning Experience</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Tue, 13 Oct 2026 03:58:40 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/csehub-turning-computer-science-articles-into-an-ai-powered-learning-experience-378b</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/csehub-turning-computer-science-articles-into-an-ai-powered-learning-experience-378b</guid>
      <description>&lt;p&gt;An open-source learning platform that combines structured technical content with article-grounded AI assistance.&lt;/p&gt;

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

&lt;p&gt;Learning computer science isn't just about finding information. It's about understanding it, connecting concepts, and getting unstuck when an explanation doesn't make sense.&lt;/p&gt;

&lt;p&gt;That's the problem I wanted to tackle with CSEHub — a computer science learning platform that brings technical articles, code snippets, and an AI learning assistant together in one place.&lt;/p&gt;

&lt;p&gt;Instead of giving learners a generic chatbot that might pull answers from anywhere, CSEHub lets users ask questions about a specific article. Its Retrieval-Augmented Generation (RAG) pipeline retrieves relevant content from that article to help ground the response in what the learner is actually studying.&lt;/p&gt;

&lt;p&gt;What can you do with CSEHub?&lt;/p&gt;

&lt;p&gt;Explore technical articles: Search, filter, and browse computer science topics, categories, and tags.&lt;/p&gt;

&lt;p&gt;Learn with code: Read structured articles with syntax-friendly code blocks and copyable snippets.&lt;/p&gt;

&lt;p&gt;Ask questions while reading: Use an AI assistant grounded in the current article, with conversation history saved for later.&lt;/p&gt;

&lt;p&gt;Manage learning content: An editor dashboard supports article creation, editing, previews, categories, tags, and publishing.&lt;/p&gt;

&lt;p&gt;Sign in securely: Google OAuth through Supabase Auth, with server-side authorization for editorial actions.&lt;/p&gt;

&lt;p&gt;Who is it for?&lt;/p&gt;

&lt;p&gt;Students preparing for placements, self-taught developers, and anyone who wants a more interactive way to learn computer science.&lt;/p&gt;

&lt;p&gt;The longer-term opportunity is to build a sustainable learning platform around high-quality technical content and AI-assisted learning. Potential business models could include premium learning tools, structured courses, or institutional access. These are future possibilities, not current revenue streams.&lt;/p&gt;

&lt;p&gt;Demo&lt;/p&gt;

&lt;p&gt;Try the live project: CSEHub&lt;/p&gt;

&lt;p&gt;Explore the learning experience and see how the platform brings technical content and AI assistance together.&lt;/p&gt;

&lt;p&gt;Code&lt;/p&gt;

&lt;p&gt;GitHub repository: github.com/captain-07/CSEHub&lt;/p&gt;

&lt;p&gt;CSEHub is released under the MIT License. The repository includes setup instructions, environment configuration, API documentation, and contribution guidelines.&lt;/p&gt;

&lt;p&gt;If you're interested in Django REST Framework, RAG pipelines, authentication, or building full-stack applications with a lightweight frontend, contributions and feedback are welcome.&lt;/p&gt;

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

&lt;p&gt;I built CSEHub with a Python backend and a static frontend, keeping the system modular so the content platform and AI assistant could evolve independently.&lt;/p&gt;

&lt;p&gt;My contribution focused on the backend and the GenAI implementation. The frontend was built with AI assistance.&lt;/p&gt;

&lt;p&gt;The technology stack&lt;/p&gt;

&lt;p&gt;Django + Django REST Framework: REST APIs, article management, validation, permissions, and user profiles.&lt;/p&gt;

&lt;p&gt;PostgreSQL: Persistent storage for articles, taxonomy, users, and conversations.&lt;/p&gt;

&lt;p&gt;Supabase Auth: Google OAuth and JWT-based authentication.&lt;/p&gt;

&lt;p&gt;LangChain + Pinecone: Retrieval pipeline and vector search for article-specific context.&lt;/p&gt;

&lt;p&gt;Google Gemini: Embeddings and AI-generated responses for the learning assistant.&lt;/p&gt;

&lt;p&gt;HTML, CSS, and JavaScript: A dependency-free static frontend.&lt;/p&gt;

&lt;p&gt;Vercel + Render: Frontend and backend deployment.&lt;/p&gt;

&lt;p&gt;How the AI assistant works&lt;/p&gt;

&lt;p&gt;The interesting part isn't simply sending a question to an LLM. It's preparing the right context before asking the model to answer.&lt;/p&gt;

&lt;p&gt;The pipeline works like this:&lt;/p&gt;

&lt;p&gt;An administrator publishes an article.&lt;/p&gt;

&lt;p&gt;The article's structured Editor.js content is converted into readable text.&lt;/p&gt;

&lt;p&gt;The text is chunked, embedded, and indexed in Pinecone.&lt;/p&gt;

&lt;p&gt;When a learner asks a question, the system retrieves relevant chunks from the article being studied.&lt;/p&gt;

&lt;p&gt;Gemini generates a response using the retrieved context.&lt;/p&gt;

&lt;p&gt;The conversation is persisted so the learner can return to it.&lt;/p&gt;

&lt;p&gt;I also kept AI indexing separate from publishing. Saving an article doesn't depend on an external embedding provider being available; indexing can be triggered separately.&lt;/p&gt;

&lt;p&gt;On the backend, I implemented server-enforced publishing permissions, token verification, content validation, and explicit handling of AI provider failures. The goal was to make the application behave like a real service, not just a demo that works when everything goes right.&lt;/p&gt;

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

&lt;p&gt;Open innovation matters when you want to build something useful without making every part of the system dependent on a single proprietary platform.&lt;/p&gt;

&lt;p&gt;CSEHub brings together open-source tools such as Django, Django REST Framework, LangChain, and the broader Python ecosystem. Their source code, documentation, and established interfaces make it possible to inspect the implementation, adapt it, and build on existing work rather than reinventing every component.&lt;/p&gt;

&lt;p&gt;There is also an important distinction in the AI layer.&lt;/p&gt;

&lt;p&gt;A closed AI API can be useful for generating responses, but the learning experience depends on more than the model itself. It depends on how content is processed, how relevant context is retrieved, how permissions are enforced, and how failures are handled.&lt;/p&gt;

&lt;p&gt;Using an open-source retrieval stack lets developers inspect and customize much of that surrounding infrastructure. The model provider can also be changed independently of the core article platform, provided the replacement supports the required interfaces.&lt;/p&gt;

&lt;p&gt;CSEHub currently uses Google Gemini, so this is not a claim that the entire AI stack is open-weight or runs locally. Instead, it demonstrates how open-source infrastructure can make an AI-powered product more modular and adaptable.&lt;/p&gt;

&lt;p&gt;For me, open innovation is about retaining the freedom to understand, modify, and extend what you build.&lt;/p&gt;

&lt;p&gt;My Agent Session&lt;/p&gt;

&lt;p&gt;Not included. I haven't attached a saved DevRelay session for this submission.&lt;/p&gt;

&lt;p&gt;Prize Categories&lt;/p&gt;

&lt;p&gt;I'm submitting CSEHub for the Side Hustle challenge. The project has a practical user problem and a potential path toward a sustainable learning product.&lt;/p&gt;

&lt;p&gt;I have not listed any additional partner prize categories because I haven't verified which ones apply.&lt;/p&gt;

&lt;p&gt;Building CSEHub taught me that adding AI to an application is the easy part to describe. Building the surrounding system — retrieval, authentication, permissions, content ingestion, persistence, and failure handling — is what turns that idea into a usable product.&lt;/p&gt;

&lt;p&gt;That's the direction I want to keep exploring: building software that solves real problems, with AI as a useful part of the experience rather than the entire product.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>Hacktoberfest Submission</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Mon, 05 Oct 2026 18:47:53 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/hacktoberfest-submission-5fik</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/hacktoberfest-submission-5fik</guid>
      <description>&lt;p&gt;What I Built&lt;br&gt;
CSEHub is a computer-science learning platform: a library of written articles with embedded code snippets, and a per-article AI assistant you can ask questions while reading.&lt;br&gt;
The problem it targets is the doomscrolling trap of learning. Most "ask AI" tools are open-ended — you type a question, get a wall of text, and twenty minutes later you're in a completely unrelated topic with nothing built. CSEHub's assistant is deliberately narrow. It's scoped to the article you're on, retrieves only chunks from that article, and is instructed to answer only from those chunks. If the answer isn't there, it says it doesn't know. You get an answer about the thing you were actually reading, you close the tab, and you've finished a topic instead of starting a scroll.&lt;br&gt;
Target users are students and early-career developers who want working explanations of data structures, algorithms, system design, and patterns, with code they can copy and run.&lt;br&gt;
Demo&lt;br&gt;
Live frontend: &lt;a href="https://cse-hub-murex.vercel.app" rel="noopener noreferrer"&gt;https://cse-hub-murex.vercel.app&lt;/a&gt;&lt;br&gt;
No video demo exists in the repository.&lt;br&gt;
Code&lt;br&gt;
&lt;/p&gt;
&lt;div class="ltag-github-readme-tag"&gt;
  &lt;div class="readme-overview"&gt;
    &lt;h2&gt;
      &lt;img src="https://assets.dev.to/assets/github-logo-5a155e1f9a670af7944dd5e12375bc76ed542ea80224905ecaf878b9157cdefc.svg" alt="GitHub logo"&gt;
      &lt;a href="https://github.com/captain-07" rel="noopener noreferrer"&gt;
        captain-07
      &lt;/a&gt; / &lt;a href="https://github.com/captain-07/CSEHub" rel="noopener noreferrer"&gt;
        CSEHub
      &lt;/a&gt;
    &lt;/h2&gt;
    &lt;h3&gt;
      
    &lt;/h3&gt;
  &lt;/div&gt;
  &lt;div class="ltag-github-body"&gt;
    
&lt;div id="readme" class="md"&gt;&lt;div class="markdown-heading"&gt;
&lt;h1 class="heading-element"&gt;CSEHub&lt;/h1&gt;
&lt;/div&gt;
&lt;p&gt;A computer-science learning platform — Django REST API + static frontend.&lt;br&gt;
Educational articles with code snippets, a per-article &lt;strong&gt;RAG chatbot&lt;/strong&gt; (Pinecone + Gemini), and &lt;strong&gt;Supabase-backed&lt;/strong&gt; user accounts.&lt;/p&gt;

&lt;p&gt;&lt;a rel="noopener noreferrer nofollow" href="https://camo.githubusercontent.com/8fd64d3c68b3007926a8aa85fc77175e719f93740e8c5f3280a1993309dbb064/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f707974686f6e2d332e31312b2d626c75653f6c6f676f3d707974686f6e"&gt;&lt;img src="https://camo.githubusercontent.com/8fd64d3c68b3007926a8aa85fc77175e719f93740e8c5f3280a1993309dbb064/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f707974686f6e2d332e31312b2d626c75653f6c6f676f3d707974686f6e" alt="Python"&gt;&lt;/a&gt;
&lt;a rel="noopener noreferrer nofollow" href="https://camo.githubusercontent.com/d72f434a58ae33578bc79f4a41d960dce5c52ade2f8c2d09f0febd4ded7c9caf/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646a616e676f2d362e302e332d3039324532303f6c6f676f3d646a616e676f"&gt;&lt;img src="https://camo.githubusercontent.com/d72f434a58ae33578bc79f4a41d960dce5c52ade2f8c2d09f0febd4ded7c9caf/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646a616e676f2d362e302e332d3039324532303f6c6f676f3d646a616e676f" alt="Django"&gt;&lt;/a&gt;
&lt;a rel="noopener noreferrer nofollow" href="https://camo.githubusercontent.com/1f8ee7d60c70d124f365a0e2db3da5996a1374184f3425b1b265d220f77f7a1b/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646a616e676f726573746672616d65776f726b2d332e31362e302d7265643f6c6f676f3d646a616e676f"&gt;&lt;img src="https://camo.githubusercontent.com/1f8ee7d60c70d124f365a0e2db3da5996a1374184f3425b1b265d220f77f7a1b/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f646a616e676f726573746672616d65776f726b2d332e31362e302d7265643f6c6f676f3d646a616e676f" alt="DRF"&gt;&lt;/a&gt;
&lt;a rel="noopener noreferrer nofollow" href="https://camo.githubusercontent.com/f8df3091bbe1149f398a5369b2c39e896766f9f6efba3477c63e9b4aa940ef14/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d677265656e"&gt;&lt;img src="https://camo.githubusercontent.com/f8df3091bbe1149f398a5369b2c39e896766f9f6efba3477c63e9b4aa940ef14/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d677265656e" alt="License"&gt;&lt;/a&gt;
&lt;a rel="noopener noreferrer nofollow" href="https://camo.githubusercontent.com/77e28b31182587815b06b5f3a017e335851f343540b277e1fa711858407deac2/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6465706c6f7965642532306f6e2d52656e6465722d3436453342373f6c6f676f3d72656e646572"&gt;&lt;img src="https://camo.githubusercontent.com/77e28b31182587815b06b5f3a017e335851f343540b277e1fa711858407deac2/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6465706c6f7965642532306f6e2d52656e6465722d3436453342373f6c6f676f3d72656e646572" alt="Render"&gt;&lt;/a&gt;&lt;/p&gt;




&lt;div class="markdown-heading"&gt;
&lt;h2 class="heading-element"&gt;Table of Contents&lt;/h2&gt;
&lt;/div&gt;

&lt;ul&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#features" rel="noopener noreferrer"&gt;Features&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#tech-stack" rel="noopener noreferrer"&gt;Tech Stack&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;
&lt;a href="https://github.com/captain-07/CSEHub#getting-started" rel="noopener noreferrer"&gt;Getting Started&lt;/a&gt;
&lt;ul&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#option-a--docker-recommended" rel="noopener noreferrer"&gt;Option A — Docker (Recommended)&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#option-b--manual-setup" rel="noopener noreferrer"&gt;Option B — Manual Setup&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;


&lt;/li&gt;

&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#environment-variables" rel="noopener noreferrer"&gt;Environment Variables&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;

&lt;a href="https://github.com/captain-07/CSEHub#usage" rel="noopener noreferrer"&gt;Usage&lt;/a&gt;

&lt;ul&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#frontend-pages" rel="noopener noreferrer"&gt;Frontend Pages&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#api-documentation" rel="noopener noreferrer"&gt;API Documentation&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#management-commands" rel="noopener noreferrer"&gt;Management Commands&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;


&lt;/li&gt;

&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#project-structure" rel="noopener noreferrer"&gt;Project Structure&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#api-endpoints" rel="noopener noreferrer"&gt;API Endpoints&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#deployment" rel="noopener noreferrer"&gt;Deployment&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;

&lt;a href="https://github.com/captain-07/CSEHub#contributing" rel="noopener noreferrer"&gt;Contributing&lt;/a&gt;

&lt;ul&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#1-fork-the-repository" rel="noopener noreferrer"&gt;1. Fork the Repository&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#2-clone-your-fork" rel="noopener noreferrer"&gt;2. Clone Your Fork&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#3-add-the-upstream-remote" rel="noopener noreferrer"&gt;3. Add the Upstream Remote&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#4-set-up-your-local-environment" rel="noopener noreferrer"&gt;4. Set Up Your Local Environment&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#5-keep-your-fork-in-sync" rel="noopener noreferrer"&gt;5. Keep Your Fork in Sync&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#6-create-a-feature-branch" rel="noopener noreferrer"&gt;6. Create a Feature Branch&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#7-make-your-changes" rel="noopener noreferrer"&gt;7. Make Your Changes&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#8-run-the-tests" rel="noopener noreferrer"&gt;8. Run the Tests&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#9-push-to-your-fork" rel="noopener noreferrer"&gt;9. Push to Your Fork&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#10-open-a-pull-request" rel="noopener noreferrer"&gt;10. Open a Pull Request&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;


&lt;/li&gt;

&lt;li&gt;&lt;a href="https://github.com/captain-07/CSEHub#license" rel="noopener noreferrer"&gt;License&lt;/a&gt;&lt;/li&gt;

&lt;/ul&gt;
&lt;br&gt;


&lt;div class="markdown-heading"&gt;
&lt;h2 class="heading-element"&gt;Features&lt;/h2&gt;
&lt;/div&gt;

&lt;p&gt;&lt;/p&gt;&lt;div class="table-wrapper-paragraph"&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;br&gt;&lt;table&gt;

&lt;thead&gt;

&lt;tr&gt;

&lt;th&gt;Feature&lt;/th&gt;

&lt;th&gt;Status&lt;/th&gt;

&lt;/tr&gt;

&lt;/thead&gt;

&lt;tbody&gt;

&lt;tr&gt;

&lt;td&gt;
&lt;br&gt;
&lt;strong&gt;Article library&lt;/strong&gt; — Browse, filter, search published articles by category/tag, with embedded code snippets&lt;/td&gt;

&lt;td&gt;✅ Mature&lt;/td&gt;

&lt;/tr&gt;

&lt;tr&gt;

&lt;td&gt;
&lt;br&gt;
&lt;strong&gt;AI article assistant&lt;/strong&gt; — Authenticated RAG chatbot (Pinecone + Gemini) scoped to a single article, with persisted conversation history&lt;/td&gt;

&lt;td&gt;✅ Implemented&lt;/td&gt;

&lt;/tr&gt;

&lt;/tbody&gt;

&lt;/table&gt;&lt;/div&gt;…&lt;p&gt;&lt;/p&gt;&lt;/div&gt;
&lt;br&gt;
  &lt;/div&gt;
&lt;br&gt;
  &lt;div class="gh-btn-container"&gt;&lt;a class="gh-btn" href="https://github.com/captain-07/CSEHub" rel="noopener noreferrer"&gt;View on GitHub&lt;/a&gt;&lt;/div&gt;
&lt;br&gt;
&lt;/div&gt;
&lt;br&gt;
&lt;br&gt;&lt;br&gt;
How I Built It&lt;br&gt;&lt;br&gt;
Two Django apps that matter: articles (the library) and chatbot (the RAG pipeline).&lt;br&gt;&lt;br&gt;
On publish, a post_save signal splits the article into chunks, embeds them, and stores them in a Pinecone namespace keyed by article_id. At question time, the flow is: similarity search filtered to that one article (k=4) → concatenate chunks → format into a system prompt → one LLM call → persist the turn in a Conversation/Message pair scoped to request.user.&lt;br&gt;&lt;br&gt;
Stack: Django 6.0.3 + DRF 3.16.0, PostgreSQL, Supabase Auth via a custom SupabaseJWTAuthentication DRF class, drf-spectacular for OpenAPI docs, static HTML/CSS/JS frontend, Docker Compose with Nginx for local, Render + Vercel for production.&lt;br&gt;&lt;br&gt;
Why Does Open Innovation Matter?&lt;br&gt;&lt;br&gt;
The honest framing: LangChain is the open-source piece, not the model. langchain-core, langchain-text-splitters, langchain-pinecone, and langchain-google-genai are all MIT-licensed and doing the real structural work here — chunking, the vectorstore abstraction, retrieval with metadata filters. The model itself is gemini-3.5-flash-lite through Gemini embeddings, which is a closed API, and the vector store is Pinecone, also closed.&lt;br&gt;&lt;br&gt;
What open source bought me was the plumbing. The RAG layer is ~150 lines of readable code instead of a vendor SDK, and because LangChain sits behind small wrappers (get_vectorstore(), get_embeddings(), get_llm()), swapping the embedding model or the vector database is a one-function change. If Pinecone's pricing or a Gemini outage breaks things, neither is welded into the architecture. A closed end-to-end RAG product would have made the ingestion and retrieval layers a black box I couldn't fork, audit, or reimplement against a local index.&lt;br&gt;&lt;br&gt;
I'm also not claiming a local-inference or open-weight setup, because there isn't one — GEMINI_API_KEY, PINECONE_API_KEY, and PINECONE_INDEX_NAME are all required in backend/.env.&lt;br&gt;&lt;br&gt;
My Agent Session&lt;br&gt;&lt;br&gt;
I don't have a DevRelay session to embed for this submission.&lt;br&gt;&lt;br&gt;
What I learned shipping it&lt;br&gt;&lt;br&gt;
I also wrote up a 50-item audit in documents/issues.md covering the backend, frontend, and deployment config — 3 critical, 7 high, 23 medium, 17 low. Some highlights: ALLOWED_HOSTS is set with an https:// scheme, which Django never matches (this is why the Render API currently500s), WhiteNoise's STATICFILES_STORAGE was removed in Django 5.1 so compression is silently off, and manage.py test reports Ran 0 tests across all four apps. Open to PRs.&lt;br&gt;&lt;br&gt;
Repository facts used

&lt;ul&gt;
&lt;li&gt;GitHub URL: &lt;a href="https://github.com/captain-07/CSEHub" rel="noopener noreferrer"&gt;https://github.com/captain-07/CSEHub&lt;/a&gt; (from git remote -v)&lt;/li&gt;
&lt;li&gt;Demo URL: &lt;a href="https://cse-hub-murex.vercel.app" rel="noopener noreferrer"&gt;https://cse-hub-murex.vercel.app&lt;/a&gt; (from backend/.env CORS_ALLOWED_ORIGINS; returns HTTP 200). Backend API &lt;a href="https://csehub-ezdl.onrender.com" rel="noopener noreferrer"&gt;https://csehub-ezdl.onrender.com&lt;/a&gt; currently returns HTTP 500 — consistent with issue C1 in documents/issues.md.&lt;/li&gt;
&lt;li&gt;AI/model/framework: LangChain (MIT, open source) — langchain-core, langchain-google-genai, langchain-pinecone, langchain-text-splitters. Model gemini-3.5-flash-lite, embeddings gemini-embedding-001, vector store Pinecone. No open-weight or local-inference model is used.&lt;/li&gt;
&lt;li&gt;Main technologies: Django 6.0.3, DRF 3.16.0, PostgreSQL, Supabase Auth (custom JWT), drf-spectacular, WhiteNoise, Gunicorn, Docker Compose + Nginx, static HTML/CSS/JS frontend, Render + Vercel.&lt;/li&gt;
&lt;li&gt;Removed: Prize Categories section — no partner categories could be verified from the repository.&lt;/li&gt;
&lt;/ul&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>Lipi Setu — Making Difficult Documents Easier to Understand</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Mon, 05 Oct 2026 06:51:27 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/lipi-setu-making-difficult-documents-easier-to-understand-1bp0</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/lipi-setu-making-difficult-documents-easier-to-understand-1bp0</guid>
      <description>&lt;p&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;/p&gt;

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

&lt;p&gt;&lt;strong&gt;Lipi Setu&lt;/strong&gt; is an AI-powered document understanding tool designed to help people understand difficult documents without needing to decode complicated language or terminology themselves.&lt;/p&gt;

&lt;p&gt;The idea came from a simple problem: important documents are often written in formal, technical, or bureaucratic language. For someone who is not comfortable with that language, understanding what the document actually says — and what they need to do next — can be difficult.&lt;/p&gt;

&lt;p&gt;Lipi Setu lets a user upload a document or image and uses multimodal AI to analyze it.&lt;/p&gt;

&lt;p&gt;Instead of simply returning a summary, it turns the document into something more actionable:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;📄 Explains what the document is about&lt;/li&gt;
&lt;li&gt;🧠 Simplifies complicated language&lt;/li&gt;
&lt;li&gt;⚠️ Identifies important information such as urgency or deadlines&lt;/li&gt;
&lt;li&gt;✅ Extracts actions the person may need to take&lt;/li&gt;
&lt;li&gt;🔊 Provides an option to listen to the explanation&lt;/li&gt;
&lt;li&gt;📝 Helps the user understand what they should do next&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The project was built with the idea of making technology useful for a real person rather than building another generic AI chatbot.&lt;/p&gt;

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

&lt;p&gt;&lt;strong&gt;GitHub Repository:&lt;/strong&gt; &lt;a href="https://github.com/captain-07/Lipi-Setu.git" rel="noopener noreferrer"&gt;https://github.com/captain-07/Lipi-Setu.git&lt;/a&gt;&lt;/p&gt;

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

&lt;p&gt;Lipi Setu is built as a lightweight Python application with a Streamlit interface.&lt;/p&gt;

&lt;p&gt;The basic workflow is:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;        Document / Image
               │
               ▼
       ┌─────────────────┐
       │   Lipi Setu UI  │
       │    (Streamlit)  │
       └────────┬────────┘
                │
                ▼
       Multimodal AI Model
                │
                ▼
       Structured Understanding
          ┌─────┼─────┐
          ▼     ▼     ▼
       Summary Actions Important Info
                │
                ▼
        Human-readable Output
                │
          ┌─────┴─────┐
          ▼           ▼
        Audio        PDF
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The project uses &lt;strong&gt;Python and Streamlit&lt;/strong&gt; for the application layer and multimodal AI for understanding uploaded documents and images.&lt;/p&gt;

&lt;p&gt;For AI, I explored Google's open model ecosystem, including &lt;strong&gt;Gemma&lt;/strong&gt;, and designed the application around multimodal document understanding rather than a traditional text-only chatbot.&lt;/p&gt;

&lt;p&gt;The AI output is structured so that the application can present useful information instead of dumping a raw model response onto the screen.&lt;/p&gt;

&lt;p&gt;I also kept the architecture intentionally simple. There is no unnecessary database or complex backend infrastructure for the core use case.&lt;/p&gt;

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

&lt;p&gt;For a project like Lipi Setu, open innovation matters because accessibility should not depend entirely on having access to a proprietary AI service.&lt;/p&gt;

&lt;p&gt;Open models such as &lt;strong&gt;Gemma&lt;/strong&gt; make it possible for developers to experiment with AI systems, understand how they behave, adapt them to specific use cases, and build applications without treating the underlying model as a completely inaccessible black box.&lt;/p&gt;

&lt;p&gt;That matters particularly for applications dealing with documents and potentially sensitive information.&lt;/p&gt;

&lt;p&gt;The long-term direction for Lipi Setu is to make the AI layer increasingly flexible — allowing the application to work with different models and potentially move more processing closer to the user instead of depending entirely on one closed provider.&lt;/p&gt;

&lt;p&gt;Open AI ecosystems make that kind of experimentation possible.&lt;/p&gt;

&lt;h2&gt;
  
  
  Prize Categories
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Build for a Friend&lt;/strong&gt;&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Open Innovation / Open Source AI&lt;/strong&gt; &lt;em&gt;(if applicable to the partner category requirements)&lt;/em&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Final Thoughts
&lt;/h2&gt;

&lt;p&gt;Lipi Setu started from a simple observation: &lt;strong&gt;understanding a document should not require understanding the language used to write it.&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The goal wasn't to build another general-purpose AI assistant. It was to build a focused tool that takes something intimidating — a complicated document — and turns it into information that a normal person can actually understand and act upon.&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>weekendchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>Highlights from Hacktoberfest Hack Day Kalyani</title>
      <dc:creator>Debjyoti Saha</dc:creator>
      <pubDate>Mon, 05 Oct 2026 06:33:52 +0000</pubDate>
      <link>https://dev.to/debjyoti_saha_63b315289e6/highlights-from-hacktoberfest-hack-day-kalyani-496k</link>
      <guid>https://dev.to/debjyoti_saha_63b315289e6/highlights-from-hacktoberfest-hack-day-kalyani-496k</guid>
      <description>&lt;p&gt;Building software alongside a community of passionate developers always hits differently. This past weekend, I attended Hacktoberfest Hack Day Kalyani, hosted by Dev Community - Kalyani Government Engineering College in collaboration with Major League Hacking (MLH).&lt;/p&gt;

&lt;p&gt;From the morning inauguration to the final demos, the energy at the Academic cum Administrative Building was incredible. Here is a quick recap of what we built, what we learned, and why events like this matter.&lt;/p&gt;

</description>
      <category>community</category>
      <category>hacktoberfest</category>
      <category>opensource</category>
      <category>programming</category>
    </item>
  </channel>
</rss>
