External data acquisition
Designed and ran a centralized ingestion platform at JPMC — product teams get licensed and open-web data without standing up their own engineering. Five formats, files to 384GB, run at 30% of its budget.
Data discovery marketplace
Built governed search across 150K+ datasets — 2M+ daily queries at sub-2-second response, metadata standardized on W3C DCAT.
Data quality orchestration
Building the orchestration layer that connects Collibra data-quality rules to the applications that consume them — runs triggered, data validated and deduped, results published over Kafka.
Data pipeline and ML infrastructure for pill identification — automated ingestion of 131K images across 4,864+ classes from NIH datasets, with format conversion, validation, and containerized deployment.
Open-source Claude Code workflow for job applications — a verified facts file is the only source of truth, so the model tailors but can never invent. Resumes, cover letters, and a personal site from three slash commands.