Technical Deep-Dive

System Architecture

Explore the technical implementation of our enterprise-grade IoT predictive maintenance platform

Repository Structure

Project Organization

predictive-care-next/

Next.js 16 Frontend (Modern UI)

Next.js 16React 19Tailwind v4Framer Motion

ml-enterprise/

Enterprise ML Backend (FastAPI + AI)

FastAPIXGBoostLightGBMCatBoostChromaDB

epics-frontend/

Original React Frontend

ReactJavaScriptCSS

epics-backend/

Original Node.js Backend

Node.jsExpressMongoDB

iot/

Arduino IoT Sensor Code

ArduinoC++DHT Sensors

ML-Model/

Original ML Notebook

JupyterPythonScikit-learn
FastAPI Backend

API Endpoints

GET/healthHealth check
POST/api/predictGet failure prediction
POST/api/predict/batchBatch predictions
POST/api/analyzeFull machine analysis
GET/api/recommendationsMaintenance recommendations
GET/api/knowledge/statsKnowledge base statistics
Model Insights

Feature Importance

Understanding what drives our ensemble model's predictions

Tool Wear35%
Temperature25%
Torque22%
Rotational Speed18%

Ensemble Model

Model TypeEnsemble (Stacking)
Base ModelsXGBoost + LightGBM + CatBoost
Top FeatureTool Wear (35%)
RAG SystemChromaDB + Sentence-Transformers
Get Started

Quick Start Guide

Prerequisites

Node.js 18+Python 3.10+PostgreSQL (optional)

1. Start the ML Backend

# Navigate to ML backend
cd ml-enterprise
# Create virtual environment
python -m venv predictive_maintenance_env
.\predictive_maintenance_env\Scripts\Activate.ps1 # Windows
# Install dependencies
pip install -r requirements.txt
# Train and serve
python main.py train --model-type ensemble
python main.py serve --port 8000

2. Start the Next.js Frontend

cd predictive-care-next
npm install
npm run dev
# Visit: http://localhost:3000

Explore the Source Code

Dive into the full implementation on GitHub and see how we built this enterprise-grade IoT platform.