<?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: HEMANG BHAVASAR</title>
    <description>The latest articles on DEV Community by HEMANG BHAVASAR (@bhavasarhemang).</description>
    <link>https://dev.to/bhavasarhemang</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.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3412149%2Fe6096c40-e7d5-477c-a4bf-ecd69ca94563.png</url>
      <title>DEV Community: HEMANG BHAVASAR</title>
      <link>https://dev.to/bhavasarhemang</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/bhavasarhemang"/>
    <language>en</language>
    <item>
      <title>Tom and Jerry Card Apps with Google AI Studio</title>
      <dc:creator>HEMANG BHAVASAR</dc:creator>
      <pubDate>Tue, 05 Aug 2025 08:42:26 +0000</pubDate>
      <link>https://dev.to/bhavasarhemang/tom-and-jerry-card-apps-with-google-ai-studio-30k8</link>
      <guid>https://dev.to/bhavasarhemang/tom-and-jerry-card-apps-with-google-ai-studio-30k8</guid>
      <description>&lt;p&gt;&lt;em&gt;This post is my submission for &lt;a href="https://dev.to/deved/build-apps-with-google-ai-studio"&gt;DEV Education Track: Build Apps with Google AI Studio&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

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

&lt;p&gt;I created the "Tom &amp;amp; Jerry Card Generator," a web application that brings the classic cartoon's chaotic energy to life. It uses the Gemini API to dynamically generate unique, vintage-style trading cards. The app leverages two different models: 'gemini-2.5-flash' to craft a witty title, a scene description, and hilarious stats, and 'imagen-3.0-generate-002' to create a matching visual in the iconic 1950s animation style. The result is an endless supply of zany cartoon moments captured on a digital card.&lt;/p&gt;

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

&lt;p&gt;When you first land on the page, you're greeted with a prompt to start the mayhem. Clicking the "Generate New Card!" button kicks off the process, calling the Gemini API to invent a brand new Tom and Jerry scenario. In seconds, a card appears with a unique image, a funny title like "Cat-apult Catastrophe," and stats for "Mischief Level," "Slapstick Rating," and "Cheese Factor." Every click generates a completely new, unpredictable, and often hilarious scene, ready to be enjoyed.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://aistudio.google.com/u/1/apps/drive/1YjPqi74Jin9xP77xuBEQoevau7wWa4ul?showPreview=true&amp;amp;resourceKey=" rel="noopener noreferrer"&gt;https://aistudio.google.com/u/1/apps/drive/1YjPqi74Jin9xP77xuBEQoevau7wWa4ul?showPreview=true&amp;amp;resourceKey=&lt;/a&gt;&lt;/p&gt;

&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.amazonaws.com%2Fuploads%2Farticles%2F1s8sahoqdqe99lxbskgm.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.amazonaws.com%2Fuploads%2Farticles%2F1s8sahoqdqe99lxbskgm.png" alt=" " width="800" height="967"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  My Experience
&lt;/h2&gt;

&lt;p&gt;Building this app was a fantastic experience, and I was impressed by how seamlessly the Google GenAI SDK integrated into the project. Using the responseSchema feature was a game-changer, as it allowed me to get perfectly structured JSON for the card's text and stats directly from the API, which eliminated the need for complex string parsing. Combining the text generation with the powerful imagen-3.0-generate-002 model made it incredibly easy to produce cohesive and visually appealing results. This project was a fun and powerful demonstration of how different AI models can work together to create engaging and creative applications.&lt;/p&gt;

</description>
      <category>deved</category>
      <category>learngoogleaistudio</category>
      <category>ai</category>
      <category>gemini</category>
    </item>
    <item>
      <title>SMART PPT AGENT</title>
      <dc:creator>HEMANG BHAVASAR</dc:creator>
      <pubDate>Mon, 04 Aug 2025 14:23:17 +0000</pubDate>
      <link>https://dev.to/bhavasarhemang/smart-ppt-agent-dm3</link>
      <guid>https://dev.to/bhavasarhemang/smart-ppt-agent-dm3</guid>
      <description>&lt;p&gt;🚀 Excited to share my latest GenAI project: SMART PPT AGENT! 🤖&lt;/p&gt;

&lt;p&gt;After 5 intensive weeks of coding, debugging, and experimenting, I've built an enterprise-grade AI-powered presentation assistant that's transforming how professionals create presentations. What an incredible learning journey it's been! Here's what makes me passionate about this space:&lt;/p&gt;

&lt;p&gt;After months of deep diving into the world of Generative AI, I've built an enterprise-grade AI-powered presentation assistant that's transforming how professionals create presentations. Here's what makes me passionate about this space:&lt;/p&gt;

&lt;p&gt;🎯 The Problem I Solved:&lt;br&gt;
Ever spent 3.5 hours converting a dense PDF report into a compelling PowerPoint presentation? I did too, until I built a solution that does it in just 15 minutes!&lt;/p&gt;

&lt;p&gt;🧠 What I Built: SMART PPT AGENT - An intelligent system that:&lt;/p&gt;

&lt;p&gt;✅ Converts ANY PDF (even scanned documents) into professional presentations&lt;br&gt;
✅ Uses Google Gemini 1.5 Flash for advanced content analysis&lt;br&gt;
✅ Preserves context and meaning while enhancing readability&lt;br&gt;
✅ Automatically highlights KPIs, financial data, and key metrics&lt;br&gt;
✅ Supports custom templates for brand consistency&lt;br&gt;
✅ Includes OCR fallback for image-based documents&lt;/p&gt;

&lt;p&gt;🔥 The Tech Stack:&lt;br&gt;
AI Engine: Google Gemini 1.5 Flash&lt;br&gt;
Frontend: Streamlit for intuitive UI&lt;br&gt;
PDF Processing: PyMuPDF + pdfplumber + Tesseract OCR&lt;br&gt;
Presentation Engine: python-pptx with smart layout selection&lt;br&gt;
Architecture: Multi-stage fallback system for 98%+ success rate&lt;/p&gt;

&lt;p&gt;📊 Real Impact:&lt;br&gt;
Time Savings: 3.5 hours → 15 minutes (95% reduction)&lt;br&gt;
Cost Efficiency: Save $150-$300 per presentation&lt;br&gt;
Accuracy: 95%+ content relevance with AI validation&lt;br&gt;
Versatility: Works across industries - from financial reports to academic papers&lt;/p&gt;

&lt;p&gt;🎓 My GenAI Learning Journey:&lt;br&gt;
This project has been incredible for understanding:&lt;br&gt;
Prompt Engineering: Crafting precise instructions for content analysis&lt;br&gt;
Multi-modal AI: Combining text extraction, image processing, and generation&lt;br&gt;
AI Agent Architecture: Building robust fallback systems and quality validation&lt;br&gt;
Context Preservation: Maintaining semantic meaning across transformations&lt;br&gt;
Production AI: Handling edge cases, error management, and user experience&lt;/p&gt;

&lt;p&gt;The GenAI space is evolving rapidly, and I'm thrilled to be building solutions that make AI accessible and practical for everyday business challenges.&lt;/p&gt;

&lt;p&gt;Who else is working on AI-powered productivity tools? Would love to connect and share experiences!&lt;/p&gt;

&lt;p&gt;Happy to connect for discussion!&lt;/p&gt;

</description>
      <category>ai</category>
      <category>python</category>
      <category>agentaichallenge</category>
      <category>learning</category>
    </item>
  </channel>
</rss>
