Data Science2024

Modular Data Science Pipeline with DVC & Dagger

Refactored a monolithic Python workflow into a modular, reproducible data science pipeline with clean separation of concerns.

Methodology

Data versioning, containerized workflow orchestration, and automated testing integrated across preprocessing, training, and model-selection stages.

Findings

Achieved consistent, traceable workflows that streamlined collaboration and supported reliable production deployment.

Tools & Methods

DVCDaggerGitHub ActionsDockerPython
Large-Scale Data Analysis and Machine Learning LifecycleElection Survey: Data Anonymization and Privacy Analysis