Umer Adam

At Cedar, I lead a team of six building a Databricks lakehouse that processes 35M events per day for analytics, product, and clinical teams. I designed the Azure Databricks medallion architecture and moved 30+ clinical, claims, and billing sources from overnight batch to Kafka, Debezium CDC, and Spark Structured Streaming.

I rebuilt 120 legacy Spark jobs as Databricks declarative pipelines, cutting runtime 55% and monthly compute cost 33%. I also established HIPAA-aligned governance through Unity Catalog, Delta Sharing, PII and PHI masking, RBAC, and column-level lineage, while building feature, embedding, vector-search, and MCP capabilities behind production models and internal LLM tooling.

Previously at Wealthfront, I led a zero-downtime migration of 400+ tables from Redshift to Snowflake, improved 95th-percentile response time 40%, and reduced Snowflake spend 30%. Across Help Scout and CoverGenius, I've built central analytics environments, reliable ETL systems, dimensional models, and data-quality practices that made reporting faster and more trusted.