AI-Driven Coffee Cultivation Support System

We are building Version 1.0 of an Agricultural AI Backend (Microservices architecture) specifically for coffee tree cultivation in Vietnam. The system requires 3 core REST APIs. All API JSON outputs must be in Vietnamese. The 3 Required APIs: Pest & Disease Detection (Computer Vision): Uses YOLOv8/EfficientNet to analyze leaf photos and identify specific coffee diseases (rust, pink disease, stem cracking) with >90% accuracy. Chemical Compliance Assistant (Vision + RAG): Takes an image of a chemical bottle, uses OCR to extract text, and uses a RAG pipeline (Langchain + Gemini/OpenAI) to check against our provided allowed/banned database. Must trigger a system action payload to "lock cultivation" if banned. Yield Prediction (Machine Learning): Uses XGBoost to process real-time IoT data (soil moisture, rainfall, etc.) and historical data to predict harvest yield. Tech Stack Requirements: Python, FastAPI YOLOv8, XGBoost LangChain, Vector DB (FAISS/ChromaDB) Docker & AWS EC2 ([login to view URL]) deployment What We Provide: Proprietary coffee disease images + public datasets CSV database of permitted/banned chemicals IoT data formats and historical yield sample data Cloud infrastructure costs (AWS/API keys are paid by us separately)

Project ID: 40366646