Alex Yu

At Amazon One Medical, I lead the architecture of a HIPAA-aligned healthcare data platform that unifies EHR/FHIR, laboratory, pharmacy, and scheduling data for approximately 500,000 patients and millions of annual encounters.

I build reusable ingestion and transformation systems with AWS, PySpark, Airflow, and Snowflake. My work reduced lab-partner onboarding from 6–8 weeks to 3–4 weeks, improved clinical terminology coverage, and reduced data latency from 6–8 hours to 45–90 minutes.

I also design patient identity-resolution services, near-real-time pipelines, and operational standards for observability, reconciliation, quarantine, and recovery. Through architecture reviews, production-readiness assessments, and mentoring, I helped reduce pipeline incidents by approximately 60% and MTTR from 90 to 35 minutes.

Previously at Databricks, I built Spark-based ETL platforms, improved reporting performance, reduced AWS compute costs from approximately $45K to $31K monthly, and strengthened production reliability through idempotent writes, validation, and automated alerts.