Secure Healthcare Data Lakehouse With Fine-Grained Access Control
Hospitals, insurers, and health networks accumulate enormous volumes of patient records, lab results, and claims data, yet most of it sits in silos that are hard to analyze and even harder to govern. Healthcare data teams need warehouse-grade reliability on inexpensive object storage, along with proof for auditors that every access to protected health information is authenticated, authorized, and logged.
The intern builds the full stack for a regulated clinical data platform: Delta Lake tables on S3 provide ACID transactions, schema enforcement, and time travel; Spark jobs written in Python move records from raw ingest through curated zones; and a Hive Metastore serves as the central schema catalog. Apache Ranger enforces fine-grained, column-level access policies, Kerberos secures cluster authentication, and masking rules redact PII before analysts ever see a row. Presto sits on top for fast ad-hoc queries, and the design is mapped explicitly to GDPR and HIPAA obligations.
Finishing the project demonstrates a rare combination of big data engineering and security governance: multi-zone lakehouse architecture, production Spark ETL, and a compliant, auditable query layer operated end to end.
