Hello! I'm Madhav Zanwar, an AI/ML student who enjoys turning ideas into working products — from real-time computer vision demos to full-stack platforms. I care most about projects that actually help people, especially around accessibility and social impact.
- 🔭 Building projects that combine AI/ML with practical, real-world use cases
- 🌱 Strengthening my fundamentals in Python, C++, DSA, and Git/GitHub
- 🧠 Focused on improving problem-solving & logical thinking
- 🤝 Open to beginner-friendly open-source contributions
🔗 github.com/madhavzanwar/kisan-sathi-app · 🌐 Live Demo Tech Stack: React.js, FastAPI, PyTorch (ResNet18), Scikit-Learn, Google Gemini 1.5 Pro, Vercel, Render
- Full-stack AI agri-tech platform engineered to empower farmers with real-time crop disease diagnosis, soil fertilizer calculation, and AI advisory
- AI Crop Diagnosis: Custom-trained PyTorch ResNet18 CNN diagnosing plant diseases from leaf images with actionable organic and chemical treatment plans
- Smart Fertilizer Calculator: Scikit-Learn Random Forest model calculating exact dosages of Urea, DAP, MOP, and Organic Compost based on soil and crop parameters
- AI Assistant: Context-aware agricultural chatbot powered by Google Gemini 1.5 Pro
- Modern glassmorphic interface built in React (Vite) with an asynchronous FastAPI backend deployed on Vercel and Render
🔗 github.com/madhavzanwar/quicksketch Tech Stack: HTML5 Canvas, CSS3, JavaScript, Flask, Flask-CORS, PyTorch, NumPy, Pillow
- Full-stack AI web app that recognizes hand-drawn sketches in real time — draw, hit Predict, get ranked results in under a second
- Custom-trained PyTorch CNN classifying across 16 categories, achieving 95.9% validation accuracy on 16,000 held-out samples
- Trained on the Google Quick Draw dataset (.npy bitmap format), returning ranked predictions with confidence scores
- Built end-to-end: HTML5 Canvas frontend + Flask/Flask-CORS backend serving the PyTorch model
🔗 github.com/madhavzanwar/placement-ai Tech Stack: Flask, HTML, CSS, JavaScript + AI integration
- Resume parsing + placement readiness scoring
- Mock interview system covering behavioral and technical rounds
- Dashboard for skill tracking and personalized roadmap
