<?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: Anmol Tiwari</title>
    <description>The latest articles on DEV Community by Anmol Tiwari (@anmoltiwari4).</description>
    <link>https://dev.to/anmoltiwari4</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%2F4157303%2F2ba81466-d23d-4c5a-9413-a530829b2e1d.jpg</url>
      <title>DEV Community: Anmol Tiwari</title>
      <link>https://dev.to/anmoltiwari4</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/anmoltiwari4"/>
    <language>en</language>
    <item>
      <title>Last Brodcast</title>
      <dc:creator>Anmol Tiwari</dc:creator>
      <pubDate>Sat, 31 Oct 2026 16:56:01 +0000</pubDate>
      <link>https://dev.to/anmoltiwari4/last-brodcast-58m3</link>
      <guid>https://dev.to/anmoltiwari4/last-brodcast-58m3</guid>
      <description>&lt;h1&gt;
  
  
  Last Broadcast — Giving History a Voice with AI
&lt;/h1&gt;

&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;p&gt;&lt;strong&gt;Last Broadcast — Bhopal: Voices Across Time&lt;/strong&gt; is an AI-powered historical archive designed to preserve, explore, and bring local stories to life.&lt;/p&gt;

&lt;p&gt;The idea came from a simple question:&lt;/p&gt;

&lt;p&gt;&lt;em&gt;What happens to a city's memories when the people who carry them are no longer around to tell their stories?&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;History is often preserved through books, documents, and photographs, but personal memories and oral histories can disappear over time.&lt;/p&gt;

&lt;p&gt;Last Broadcast explores how AI can help preserve these fragments of collective memory.&lt;/p&gt;

&lt;p&gt;The platform aims to let users:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Explore curated historical collections about Bhopal.&lt;/li&gt;
&lt;li&gt;Upload audio recordings and convert speech into searchable text.&lt;/li&gt;
&lt;li&gt;Extract meaningful information from historical narratives.&lt;/li&gt;
&lt;li&gt;Ask questions about archived material using AI-powered retrieval.&lt;/li&gt;
&lt;li&gt;Listen to AI-generated narration of historical content.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I interpreted &lt;strong&gt;Ghost in the Machine&lt;/strong&gt; as the idea of giving a voice to memories that might otherwise remain buried in documents, recordings, and forgotten stories.&lt;/p&gt;

&lt;p&gt;The goal isn't to replace human memories with AI-generated history. It's to build a tool that helps preserve, organize, and make those memories more accessible.&lt;/p&gt;

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

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Live Application:&lt;/strong&gt; &lt;a href="https://last-brodcast-web.onrender.com" rel="noopener noreferrer"&gt;Last Broadcast — Add your deployed frontend URL&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Backend API:&lt;/strong&gt; &lt;a href="https://last-broadcast-api.onrender.com" rel="noopener noreferrer"&gt;https://last-broadcast-api.onrender.com&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;API Documentation:&lt;/strong&gt; &lt;a href="https://last-broadcast-api.onrender.com/docs" rel="noopener noreferrer"&gt;https://last-broadcast-api.onrender.com/docs&lt;/a&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Health Check:&lt;/strong&gt; &lt;a href="https://last-broadcast-api.onrender.com/health" rel="noopener noreferrer"&gt;https://last-broadcast-api.onrender.com/health&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&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/AnmolTiwari4/Last-Brodcast" rel="noopener noreferrer"&gt;Last Broadcast — Source Code&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The project is built with a modular architecture, separating the frontend, backend, AI services, and data layer.&lt;/p&gt;

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

&lt;p&gt;I wanted to build something that combined open-source AI with a practical application rather than simply wrapping an LLM in a chatbot interface.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Tech stack:&lt;/strong&gt;&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Frontend:&lt;/strong&gt; Next.js, TypeScript, Tailwind CSS and shadcn/ui.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Backend:&lt;/strong&gt; Python, FastAPI, SQLAlchemy and Alembic.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Database:&lt;/strong&gt; PostgreSQL with pgvector for vector-based retrieval.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;LLM:&lt;/strong&gt; Qwen 2.5 7B Instruct through Hugging Face Inference Providers.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Embeddings:&lt;/strong&gt; &lt;code&gt;sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Speech-to-Text:&lt;/strong&gt; ElevenLabs Scribe.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Text-to-Speech:&lt;/strong&gt; ElevenLabs multilingual TTS.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Deployment:&lt;/strong&gt; Render.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The intended AI workflow is:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Audio recordings are submitted to the platform.&lt;/li&gt;
&lt;li&gt;Speech-to-text converts recordings into text.&lt;/li&gt;
&lt;li&gt;The resulting content is organized and processed into searchable information.&lt;/li&gt;
&lt;li&gt;Embeddings enable semantic retrieval from the archive.&lt;/li&gt;
&lt;li&gt;Qwen generates responses grounded in retrieved historical material.&lt;/li&gt;
&lt;li&gt;ElevenLabs provides audio narration for an accessible listening experience.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;I also wanted the architecture to remain modular, so individual AI components can be replaced without rebuilding the entire application.&lt;/p&gt;

&lt;p&gt;An important design principle is distinguishing verified historical information, personal testimony, editorial interpretation&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>A Coach for students to organize their tasks well</title>
      <dc:creator>Anmol Tiwari</dc:creator>
      <pubDate>Sun, 04 Oct 2026 14:54:32 +0000</pubDate>
      <link>https://dev.to/anmoltiwari4/a-coach-for-students-to-organize-their-tasks-well-3i5c</link>
      <guid>https://dev.to/anmoltiwari4/a-coach-for-students-to-organize-their-tasks-well-3i5c</guid>
      <description>&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fccs5g1evqmgf9a40bp7w.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fccs5g1evqmgf9a40bp7w.png" alt=" " width="800" height="426"&gt;&lt;/a&gt;&lt;br&gt;
_This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend&lt;br&gt;
_&lt;br&gt;
&lt;strong&gt;What I Built&lt;/strong&gt;&lt;br&gt;
I built GoalSync AI, a personalized AI goal and progress coach created specifically for a friend who struggled with staying accountable to long-term career goals while managing daily tasks.&lt;/p&gt;

&lt;p&gt;It acts as an interactive daily check-in companion that breaks down big long-term visions into realistic weekly milestones and concrete daily to-do items. Instead of being just another static task manager, GoalSync AI chats with my friend every evening, reviews completed tasks, checks alignment with weekly milestones, and provides encouraging, structured progress recaps.&lt;/p&gt;

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

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Live Flowise Share Link:&lt;/strong&gt; &lt;a href="https://jolt-dreamt-exposable.ngrok-free.dev/chatbot/07c44f12-d49e-4dbb-a693-cd00ffab3fa7" rel="noopener noreferrer"&gt;GoalSync AI Live Demo&lt;/a&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;&lt;strong&gt;Visual Workflow &amp;amp; Demo:&lt;/strong&gt;&lt;br&gt;
The AI agent in action: chatting with the user, evaluating daily tasks against weekly milestones, and holding a contextual multi-turn conversation.&lt;br&gt;
&lt;strong&gt;Visual Workflow &amp;amp; Demo:&lt;/strong&gt;&lt;br&gt;
The AI agent in action: chatting with the user, evaluating daily tasks against weekly milestones, and holding a contextual multi-turn conversation.&lt;/p&gt;

&lt;p&gt;Code&lt;br&gt;
The entire agent flow can be exported and imported into any self-hosted or local Flowise AI instance.&lt;/p&gt;

&lt;p&gt;How I Built It&lt;br&gt;
GoalSync AI is built entirely on top of an open-source visual agent framework paired with fast open-weight model inference:&lt;/p&gt;

&lt;p&gt;Agent Harness / Framework: Flowise AI — An open-source drag-and-drop framework used to connect LLM nodes, memory blocks, and prompt templates into an agentic workflow.&lt;/p&gt;

&lt;p&gt;Open-Weight LLM Core: Hosted via Groq Cloud for high-speed inference. We leveraged open-weight models like Meta's Llama 3 to handle reasoning, planning, and progress tracking.&lt;/p&gt;

&lt;p&gt;Contextual Memory: Integrated a Buffer Memory node linked to a Conversational Chain node to ensure the agent remembers multi-turn chat context across daily check-ins.&lt;/p&gt;

&lt;p&gt;Structured Coaching Prompt: Configured system instructions directly inside the flow parameters to enforce an encouraging, structured response style with daily/weekly/long-term tracking logic.&lt;/p&gt;

&lt;p&gt;Why Does Open Innovation Matter?&lt;br&gt;
Open innovation was the core reason GoalSync AI was possible to build and deliver effectively to my friend:&lt;/p&gt;

&lt;p&gt;Data Privacy &amp;amp; Personal Control: Daily personal goals, habits, and career targets are sensitive. Using an open-source framework ensures that user data and interaction logic stay strictly within controlled pipelines rather than being funneled into proprietary training sets.&lt;/p&gt;

&lt;p&gt;Zero Cost to Run: Proprietary API tokens can quickly become expensive for everyday conversational bots. Leveraging open-weight models via high-speed inference infrastructure means my friend gets a powerful tool completely free of cost.&lt;/p&gt;

&lt;p&gt;Model Flexibility &amp;amp; No Vendor Lock-in: If a specific model endpoint changes or depreciates, the visual open canvas allows us to swap models (e.g., from Llama 3 to Mixtral) or adjust prompt parameters in seconds without rewriting a single line of application code.&lt;/p&gt;

&lt;p&gt;Prize Categories&lt;br&gt;
Main Category: Build for a Friend&lt;/p&gt;

&lt;p&gt;Tags: #hf26challenge, #devchallenge, #weekendchallenge, #ai&lt;/p&gt;

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