CloakData

Operational datasets often contain personally identifiable or sensitive values, but anonymization steps are commonly written as one-off scripts. That makes pipelines harder to audit, slower to reuse, and risky to evolve.

The project demonstrates the kind of engineering I value: privacy-aware data tooling, explicit operational feedback, and practical abstractions that can move from local experimentation into repeatable ETL work.

• A Python package focused on anonymization workflows using Polars as the transformation engine.
• Configurable rules that make masking behavior explicit instead of hidden inside ad hoc scripts.
• A library shape suitable for ETL pipelines, notebooks, and local validation workflows.

• Polars-first transformations to keep data processing fast and expressive.
• Rule-oriented design so anonymization policies can be reviewed and changed deliberately.
• Packaging for PyPI to make installation and reuse straightforward across projects.