BroiLabs

Project Overview

BroiLabs is an intelligent poultry farm monitoring platform that integrates Machine Learning models (Random Forest and XGBoost) for real-time environmental prediction and anomaly detection. The system monitors critical sensor data including temperature, humidity, and ammonia levels inside the chicken coop, while providing predictive analytics to optimize farm conditions. It also includes comprehensive production management features such as feed tracking, water consumption, population records, and mortality logging, all updated in real-time via WebSocket connections.

The Challenge

• Designing a full-stack system that seamlessly integrates real-time IoT sensor data streaming with Machine Learning prediction models, while maintaining low-latency updates across the dashboard
• The platform needed to handle continuous data ingestion from multiple sensors, run ML inference on incoming data, and present actionable insights to farm operators without delays or data loss

The Solution

• Built a high-performance backend using FastAPI with WebSocket support for real-time bidirectional communication between sensors and the frontend dashboard
• Implemented Random Forest and XGBoost models for environmental condition prediction and anomaly detection, served through dedicated ML endpoints
• Designed the frontend with Next
• js for a responsive, real-time dashboard displaying live sensor readings, prediction results, and production management tools
• Used PostgreSQL for persistent storage of historical sensor data, production records, and ML model outputs