A/B Testing in Production with Feedback Logging | Real-Time Demo | For Beginners
GitHub Repo: https://github.com/iQuantC/AB_Testing...
Support Us: https://buymeacoffee.com/iquantconsult
📝 Description:
What happens when you pit two machine learning models against each other — in real-time?
In this hands-on MLOps project, I walk you through how to deploy, test, and compare two competing ML models using:
1. Docker & Flask
2. NGINX for A/B traffic splitting
3. User feedback logging for real-world evaluation
4. Streamlit dashboard for visual insights
5. Fully containerized with Docker Compose!
🎯 Whether you're a Data Scientist, ML or MLOps beginner or looking to sharpen your deployment skills, this video shows you how to:
1. Serve multiple ML models in production
2. Route traffic between them intelligently
3. Collect and analyze feedback from live predictions
4. Make data-driven decisions about which model performs better
🔧 Tools:
Python, Flask, Scikit-learn, Docker, NGINX, Streamlit, Pandas
💬 Drop a comment if you'd like to see this deployed on the cloud or integrated with a database!
🔔 Don’t forget to Like, Comment, and Subscribe for more real-world ML and MLOps projects.
⏱️ Timestamps:
0:00 - Intro
02:50 - Project Setup
06:39 - Model Training
08:42 - Serve ML Models with Flask API Endpoints
11:51 - Route Traffic between ML Models with Nginx
14:58 - Containerize ML Models & Nginx Load Balancer with Docker & Docker Compose
23:48 - User Feedback Logging for Real-world Evaluation
27:02 - Analyze Feedback from Live Predictions
30:09 - Create interactive Dashboard with Streamlit
31:46 - Final Recap & Clean Up
#abtesting #MLOps #datascience #Nginx #docker #Streamlit #MachineLearning #RealTimeML #Python #AIProjects
Disclaimer: This video is for educational purposes only. The tools and technologies demonstrated are subject to change, and viewers are encouraged to refer to the official documentation for the most up-to-date information.
Happy MLOpsing! 🎉