🚀 Vision AI demo with Google’s Gemma 4 Over the past few days since its release, I’ve been exploring the vision capabilities of Gemma 4. I’m sharing a Streamlit web app where you can upload images and ask natural language questions about them — all running locally via Ollama. ✨ What it does • Analyze any image (JPG, PNG, GIF, etc.) • Ask questions in natural language • Get AI-powered answers using gemma4:e2b (2B, 4-bit quantized) • Runs entirely on your machine — no cloud APIs 🛠️ Tech stack • Python • Streamlit • Ollama • Gemma 4 💻 The project includes: • Web UI (Streamlit) • CLI demo • VS Code debug setup 👉 Check it out: https://lnkd.in/dtegduUc #Gemma4 #AI #MachineLearning #ComputerVision #LocalAI #Ollama #Streamlit #Python #OpenSource

Have you ever stared at a massive PDF or textbook and thought, "I wish I could just ask this book a question"? 🤔 That exact problem led me to build BookBuddy, my personal AI reading assistant, and today I’m making it open-source! I wanted a tool that didn't just understand documents, but could manage multiple books at once, keep their contexts perfectly isolated, and—most importantly—keep my personal data completely private. With BookBuddy, you can: ✅ Upload multiple PDFs and switch between them using a sleek, tabbed UI. ✅ Maintain separate, isolated chat memories for each document. ✅ Run everything 100% locally. No API keys, no cloud servers, total privacy. I built the backend using Python, FastAPI, and LangChain, utilising ChromaDB for vector storage and Ollama (Llama 3) for local inference. The frontend is a custom-designed React application built for speed and aesthetics. I’ve structured the repo so anyone can clone and launch the entire full-stack application with just one `make run` command. 🔗 Check out the repository here: https://lnkd.in/dKT8MpTQ If you give it a try, let me know! Feedback and contributions are always welcome. 🚀 #SoftwareEngineering #AI #LocalLLM #WebDevelopment #React #FastAPI #LangChain #Python

Building a Hybrid Scoring Engine for Food Transparency 🚀 I just wrapped up a major update to SmartLabel-AI, a full-stack food ingredient scanner. The biggest challenge? Converting messy real-world OCR data into accurate health insights. The Tech Stack: 🔹 Backend: FastAPI & Python 🔹 Computer Vision: Pytesseract + OpenCV with a custom ocr_correction pipeline using RapidFuzz for dictionary-based matching. 🔹 The "Brain": A hybrid ML engine. It blends a trained Ridge Regression model with rule-based heuristics. It even accounts for logarithmic positional weighting (because the first ingredient matters more than the last!). 🔹 Frontend: React & Vanilla CSS for a custom, high-performance UI. 🔹 Experience: Interactive background scenes and a seamless theme-toggle system. This project pushed me to think about data reliability and user experience in a way that feels premium and fast. Check out the repo here: https://lnkd.in/eNAJqdeW

🚀 𝗘𝘅𝗰𝗶𝘁𝗲𝗱 𝘁𝗼 𝗦𝗵𝗮𝗿𝗲 𝗠𝘆 𝗣𝗿𝗼𝗷𝗲𝗰𝘁: 𝗟𝗼𝗮𝗻 𝗔𝗽𝗽𝗿𝗼𝘃𝗮𝗹 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻 𝗦𝘆𝘀𝘁𝗲𝗺 I’m happy to present my latest project – a Loan Approval Prediction System built using Machine Learning and Flask API. 💡 𝗪𝗵𝗮𝘁 𝘁𝗵𝗶𝘀 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝗱𝗼𝗲𝘀: This system predicts whether a loan application will be Approved ✅ or Rejected ❌ based on key factors like: CIBIL Score Annual Income Number of Dependents Other financial details ⚙️ 𝗞𝗲𝘆 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀: Real-time prediction using a trained ML model Simple and user-friendly interface Backend powered by Flask API Clear decision output with probability insights 🧠 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝗶𝗲𝘀 𝗨𝘀𝗲𝗱: Python Machine Learning (Model Training) Flask (API Development) HTML/CSS (Frontend UI) 📊 𝗪𝗵𝗮𝘁 𝗜 𝗹𝗲𝗮𝗿𝗻𝗲𝗱: End-to-end ML project development Model training and evaluation API integration with frontend Handling real-world financial datasets This project gave me hands-on experience in building a complete 𝗠𝗟-𝗽𝗼𝘄𝗲𝗿𝗲𝗱 𝘄𝗲𝗯 𝗮𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻. 🎯 𝗟𝗼𝗼𝗸𝗶𝗻𝗴 𝗳𝗼𝗿𝘄𝗮𝗿𝗱 𝘁𝗼 𝗶𝗺𝗽𝗿𝗼𝘃𝗶𝗻𝗴 𝘁𝗵𝗶𝘀 𝗳𝘂𝗿𝘁𝗵𝗲𝗿 𝗮𝗻𝗱 𝗲𝘅𝗽𝗹𝗼𝗿𝗶𝗻𝗴 𝗺𝗼𝗿𝗲 𝗿𝗲𝗮𝗹-𝘄𝗼𝗿𝗹𝗱 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀! #MachineLearning #Python #Flask #DataScience #AI #WebDevelopment #Projects #LearningJourney

I got tired of manually clicking through my app after every change. You build a feature… and then spend the next 10–15 minutes just verifying it works. No proper tests. Just clicking around. So I built something that does it for me. Introducing RunSight. npx runsight ./my-project It: • Detects your project (Node.js, Python, or even static HTML) • Starts your dev server automatically • Launches a browser and explores your UI • Uses AI vision to decide what to click • Captures screenshots at every step • Records a full demo video • Outputs a structured report.json All in seconds. The part I’m most excited about: It works as a skill for Kiro, Cursor, and Claude - so your AI coding assistant can run it and understand your app without you explaining anything. This is v1 - just shipped, already usable, and I’m actively improving it based on feedback. I’m exploring a future where your project can demo itself. Would love feedback from builders here 👇 📦 npm: https://lnkd.in/gYhJj9YH ⭐ GitHub: https://lnkd.in/gEFV3_Mi #buildinpublic #opensource #devtools #ai #nodejs

🚀 House Price Prediction Web Application 🏠 I’m excited to share my recent project where I built a Machine Learning-based web application to predict house prices based on user inputs. The application allows users to enter details such as location, square footage, number of bedrooms (BHK), and bathrooms, and instantly provides an estimated property price through a trained model. ⚙️ Tech Stack: Machine Learning: Linear Regression Library: Scikit-Learn Backend: Flask (Python) Frontend: HTML, CSS, JavaScript 📊 Model Performance: The model achieved an accuracy of approximately 85–90% (R² score) after applying data cleaning, feature engineering, and outlier handling techniques. 🧠 Key Learnings and takeaways: Deploying machine learning models using Flask Importance of data preprocessing and feature engineering Integrating frontend with backend APIs Building end-to-end ML applications 🔗 GitHub Repository: https://lnkd.in/g-Cx8eqz Ongoing efforts are focused on strengthening my understanding of machine learning and improving prediction accuracy. #MachineLearning #WebDevelopment #Flask #DataScience #Python

🚀 I'm excited to share my latest project: a full‑stack **Vehicle Tracking System** with integrated AI/ML models – built from scratch and now live on GitHub. 🔧 **Tech Stack**: - Backend: FastAPI, Python, WebSockets, scikit‑learn - Frontend: React, TypeScript, Tailwind CSS, Leaflet, Recharts - Database: SQLite with Alembic migrations 🤖 **AI/ML Features**: ✅ ETA prediction using regression ✅ Anomaly detection (speeding, harsh braking) ✅ Driver behavior scoring (Eco / Normal / Aggressive) 🗺️ **Real‑time Capabilities**: - Live vehicle positions on interactive map - WebSocket updates every 2‑5 sec - Instant alerts for geofence breaches & over‑speeding 📊 **Other Highlights**: - Monthly PDF reports - Dark/light mode - Multi‑user authentication & vehicle ownership - Dynamic trip lifecycle (auto‑start/end) This project was a deep dive into real‑time systems, full‑stack integration, and practical ML deployment. 👨💻 **GitHub Repo**:   🔗 https://lnkd.in/dsx5vY9J Would love to hear your feedback or ideas for improvement! 🙌 #VehicleTracking #AI #MachineLearning #FastAPI #React #FullStack #RealTime #Python

Every developer who has built a web scraper knows this pain: Your scraper works perfectly. Then the website moves one <div>. And everything breaks. Welcome to the endless loop of fixing selectors. A new Python tool called Scrapling is getting a lot of attention for exactly this reason. Instead of relying only on CSS selectors… it “remembers” the element. Scrapling stores a fingerprint of the element its tag, attributes, neighbors, and structure. So when a website layout changes… it can relocate the element automatically. Meaning your scraper doesn’t instantly explode every time a site sneezes. It also packs some surprisingly powerful features: – Stealth fetchers to avoid bot detection – Built-in proxy rotation – Async spiders for large crawls – Browser fetchers when JavaScript rendering is needed Basically: BeautifulSoup simplicity Scrapy style crawling Playwright level dynamic fetching All in one library. That’s why developers are suddenly paying attention. Because most scraping projects don’t fail from scale… They fail from maintenance. The real cost of scraping isn’t writing the scraper. It’s fixing it every time the page changes. Tools like this shift scraping from: “babysitting fragile scripts” to “running resilient data pipelines.” Curious how many data teams are still maintaining broken scrapers every week. How do you handle scraper maintenance today? #Python #WebScraping #DataEngineering #AI #OpenSource

🔍 Showcasing my Friend Suggestion System I previously built an ML-powered friend recommendation system, and I’ve now deployed it to demonstrate how mutual connections and user interactions can be used to generate meaningful friend suggestions. 🔍 How it works: The system analyzes user data such as shared interests, activity patterns, and existing connections. Using similarity-based algorithms, it ranks and recommends relevant profiles. To make the concept more intuitive, I demonstrated the working using a graph-based approach (e.g., Alice → Bob → Charlie → David → Alice), where users are represented as nodes and connections as edges. This allows anyone to simulate and understand the backend process of mutual friend connections. 💡 Key Highlights: • Achieved 88% accuracy in friend recommendations • Improved user engagement by 30% • Delivered 92% user satisfaction in internal testing • Clean and intuitive user interface ⚙️ Tech Stack: Python, C++, HTML, CSS, JavaScript, File Handling,cloud hosting,Graph algorithms. ☁️ Deployment: Deployed on Render, making the system accessible for real-time simulation and better understanding of the recommendation logic. 🎥 Demo Video: In the video below, I demonstrate how the system works using a graph example to simulate real-world friend connections. 🔗 Live Demo:https://lnkd.in/gbXmGTEn 💻 GitHub Repository: https://lnkd.in/gZFNQ9u4 This deployment helped me showcase how recommendation systems and graph-based logic work together in real-world applications. Would love your feedback! 🙌 #MachineLearning #GraphTheory #WebDevelopment #CloudComputing #Render #Python #StudentDeveloper