Projects
Working code, not just papers. These are the implementations behind the research, plus the tools I've built around securing and modernizing data in regulated environments.
Security & privacy research
pymayfly
Identity-Per-Transaction credential library for regulated data pipelines. Like a mayfly, each credential lives for exactly one transaction, then it's gone. Open-source with AWS, Azure, and GCP backends, published to PyPI. This is the implementation behind the IEEE BigDataSecurity reference architecture.
Ghost Drivers
The code behind the Ghost Drivers research: a multi-pass spatiotemporal sieve that re-identifies "anonymized" fleet telemetry with 96% precision. Ships with a synthetic dataset so the result is reproducible without exposing real telemetry.
Legacy modernization
CobolBreaker
An LLM-assisted ETL pipeline that ingests legacy COBOL source, analyzes its structure, and generates modern documentation plus Java and Python migration scaffolding. Aimed at the modernization problem that regulated industries (finance, government, insurance) actually face: decades of COBOL nobody wants to touch.
Graph & Data Lineage
Crypto Journey
An educational blockchain transaction-lineage explorer that traces a cryptocurrency address's history, deployed as a serverless API. A hands-on take on data provenance and lineage tracing, which is the same problem space as tracking authorization through a data pipeline.
More
The full portfolio, including graph-analytics and data-engineering side projects, lives on my personal site. For the research these projects support, see the research page.