Overview
An earlier machine learning deployment project for house price prediction, built as a Flask application with an AWS deployment workflow and automated GitHub Actions pipeline.
Architecture
- Flask application — serves the prediction interface and inference flow
- AWS ECR — stores the Docker image built by the CI pipeline
- AWS EC2 — hosts the deployed container and self-hosted GitHub runner
- GitHub Actions — runs CI and pushes the deployment image
Features
- Interactive prediction UI — users can change housing features and see updated predictions
- Containerized deployment workflow — Docker image built and pushed through GitHub Actions
- End-to-end project structure — data ingestion, transformation, training, and prediction pipelines in one codebase
Tech Stack
- Python (Scikit-learn, Pandas, Flask)
- AWS (ECR, EC2)
- GitHub Actions
- Docker
Links
Key Takeaway
This project was an early end-to-end ML deployment exercise focused on packaging a prediction model as a containerized web app with an AWS deployment workflow and CI/CD automation.