PROJECT
Developed a robust machine learning pipeline to predict the likelihood of heart disease in patients leveraging MLOPs practices
Utilized Pandas for efficient data handling, cleaning, and preparation, ensuring high-quality datasets for model training.
Employed Scikit-learn for building and evaluating predictive models, experimenting with different algorithms to achieve optimal prediction accuracy.
Used Matplotlib for creating insightful visualizations, facilitating a deeper understanding of the data and model performance.
Containerized the application with Docker, achieving a seamless and consistent execution environment across development, testing, and production stages, thereby minimizing ”it works on my machine” issues.
Managed the complete machine learning lifecycle with MLflow, covering experiment tracking, model versioning, reproducibility, and deployment.
Orchestrated the entire workflow, from data ingestion and transformation to model training and exporting, utilizing Mage.
Deployed the model for batch predictions using Docker and for real-time predictions using Gunicorn, Flask, and Docker.
Incorporated Evidently for continuous model monitoring, generating detailed reports on model performance and data drift, which facilitated proactive adjustments to maintain high prediction accuracy.
Adopted industry best practices, including the implementation of unit tests to ensure code quality and reliability, and the use of Makefile for automating routine tasks and simplifying project setup and execution.