Overview
What this is
The compact read before the technical details.
An end-to-end agritech platform with RAG chatbot, crop disease detection, personalized recommendations, and live price tracking. Built with PyTorch, LangChain, and FastAPI, it combines a fine-tuned disease classification model, a RAG pipeline for farming advice, scikit-learn-based crop recommendation, and web-scraped commodity prices. Deployed with Docker for easy scaling across rural connectivity constraints.
Capabilities
What it actually does
The useful parts, pulled out of the paragraph wall.
End-to-end agritech platform combining multiple AI pipelines: RAG-based chatbot for agricultural Q&A, CNN-based crop disease detection from leaf images, and scikit-learn models for personalized crop recommendations
Live price tracking system built with BeautifulSoup and Selenium scrapers that ingest real-time mandi (market) price data, enabling farmers to make informed selling decisions
Full-stack deployment with FastAPI serving inference endpoints, Flask for the web interface, and Docker containerization for reproducible multi-service orchestration
Implementation
Technology with jobs attached
Names are less useful than responsibilities. This is what each piece is doing.
PyTorch
Deep learning framework for training and serving the crop disease detection CNN model
LangChain
RAG orchestration — retrieval-augmented generation pipeline for the agricultural chatbot
Scikit-learn
Classical ML models for crop recommendation based on soil parameters, climate data, and historical yield
FastAPI
High-performance async API serving inference endpoints for disease detection and recommendation models
Flask
Lightweight web framework powering the user-facing dashboard and chatbot interface
BeautifulSoup + Selenium
Web scraping stack for extracting live crop prices from government mandi databases
Docker
Containerization for multi-service deployment, isolating the inference API, scraping workers, and web server