Krishivani: AI-Powered Plant Disease Detection and Market Intelligence Platform
Built an AI-powered plant disease detection and market intelligence platform using Flutter, Python, FastAPI, TensorFlow/Keras, Supabase, PostgreSQL, Generative AI, and deep learning. Trained an EfficientNetB0 CNN on 54,000 images across 38 disease classes, achieving 98% accuracy, precision, and recall on held-out test data. Deployed FastAPI inference integrated with Flutter, designed a Supabase/PostgreSQL backend with row-level security, authentication and diagnosis records, and implemented a RAG-based agriculture chatbot and market-intelligence module.
LoopedIn: Sustainable Second-Hand Fashion Exchange App
Developed a cross-platform marketplace using Flutter, Supabase, Figma, PostgreSQL, REST APIs, FastAPI, and Scikit-learn. Used Clean Architecture and Riverpod in a feature-first modular codebase with 20+ production-ready screens. Built a content-based recommendation system with automated preprocessing pipelines and four product attribute vectors, and integrated authentication, PostgreSQL, REST APIs, and real-time backend services.
KrishiVani
Created a complete UI/UX design in Figma for Krishivani, an app that empowers farmers to detect plant diseases using camera scans, symptom listing, and voice input. Designed features like market price prediction and chatbot assistance to provide farmers with actionable insights. Focused on minimal, accessible, and farmer-friendly design, ensuring smooth usability in rural contexts.
LoopedIn
LoopedIn is a sustainable second hand fashion exchange project. Helps users to buy, rent or sell items. It promotes circular fashion and helps reduce waste while giving users a rewarding, trustworthy experience. Find app here: https://github.com/kanu-3/LoopedIn_v2
