mlops pipeline

project overview: a production-ready mlops system with automated ci/cd pipelines for three healthcare prediction models. the project includes a fastapi backend, streamlit frontend, dockerized deployment, and github actions automation, demonstrating an end-to-end ml workflow from model development to production. technical highlights: implemented mlflow for experiment tracking and model registry, with xgboost models trained using smote to handle class imbalance and shap for model explainability. the system follows a modular architecture with containerization, automated deployments, and cold-start optimizations, addressing real-world production and scalability challenges. impact: this project showcases my ability to build, deploy, and maintain complete ml products not just train models. it reflects hands-on experience across data science and engineering, making it well suited for ml engineer and mlops roles that require ownership of the full ml lifecycle