Skip to content

An end-to-end MLOps pipeline for network intrusion detection using Airflow, Docker, and MLflow. CyberFlow automates data ingestion, drift detection, A/B model testing, and deployment, ensuring real-time monitoring and high accuracy.

Notifications You must be signed in to change notification settings

DevManpreet5/network_Security_end2toend

Repository files navigation

CyberFlow

An end-to-end MLOps pipeline for network intrusion detection using Airflow, Docker, and MLflow. CyberFlow automates data ingestion, drift detection, A/B model testing, and deployment, ensuring real-time monitoring and high accuracy.

Features

Dual-Pipeline System

  • One-Time Training Pipeline: Trains the initial model.

  • Weekly Ingestion Pipeline: Handles new data while maintaining 98.89% accuracy.

    Automated Drift Detection

  • Weekly drift detection with detailed reports and visualizations.

  • Triggers model retraining if significant drift is detected.

    A/B Testing & Deployment

  • Compares multiple models in production.

  • Automatically deploys the best-performing model.

Dashboard for Monitoring

  • Real-time performance visualization.
  • Model metrics and drift detection reports.

Visualization

Streamlit Report: View Dashboard

Airflow DAG Running

Airflow DAG Running 2

MLFlow UI

Tech Stack

  • Python
  • Apache Airflow
  • Docker
  • MLflow
  • Streamlit

About

An end-to-end MLOps pipeline for network intrusion detection using Airflow, Docker, and MLflow. CyberFlow automates data ingestion, drift detection, A/B model testing, and deployment, ensuring real-time monitoring and high accuracy.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published