NIST Genesis is a service-oriented anomaly detection API for predictive maintenance of industrial equipment. Designed to transform exploratory machine learning notebooks into production-ready API services, it implements robust, runtime-configured inference and operational monitoring workflows.
- FastAPI Service Architecture: Clean API endpoints for real-time inference, model health checks, and service metrics, built to support cloud-ready environments.
- Runtime Inference Configuration: Decoupled, explicit model configuration allowing seamless updates to the champion model without code changes or hardcoded notebook dependencies.
- Idempotent Webhook Integration: A robust ERP/CMMS webhook integration pattern that avoids duplicate maintenance work orders from redundant device-timestamp alerts.
- Production Monitoring: Pre-configured Prometheus-style operational metrics, health endpoints, and system status checks for model deployment evaluation.
- Cloud Deployment Playbook: Optimized configuration and documentation for cost-effective deployment on AWS Lightsail or lightweight container environments.
Tech Stack
- API Framework: FastAPI, Uvicorn
- Machine Learning: scikit-learn
- Data Science: NumPy, Pandas, Jupyter Notebook
- DevOps & Infrastructure: Docker, Docker Compose, Prometheus (Metrics)
- Deployment: AWS Lightsail / Cloud-ready configurations
Installation
- Clone repository:
git clone https://github.com/JoshuaGlaZ/nist-genesis.git
cd nist-genesis
- Set up virtual environment and install dependencies:
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txt
- Start the FastAPI application:
uvicorn app.main:app --reload