Diego Costantino

I build and operate governed lakehouse platforms on Databricks and AWS, using Delta Lake, PySpark, Structured Streaming, CDC, Unity Catalog, Terraform, and CI/CD.

At NapoliData, I orchestrate Databricks Workflows across Delta Lake medallion layers, deploy reusable AWS Glue and Step Functions infrastructure, and operate Apache Iceberg pipelines with automated data validation. I also built a LangGraph and Amazon Bedrock observability layer and a pgvector RAG solution focused on inference cost and latency control.

Across Proactiviti, Fivvy, Aprende Institute, and Johnson & Johnson, I've delivered cross-cloud ingestion, Airflow-on-Kubernetes orchestration, AWS migrations, Redshift performance improvements, and production ML pipelines. My recent Snowflake and dbt work supports multi-tenant platforms with more than 100 data assets serving roughly 50 tenants.

My background spans 17+ years in data, analytics, and risk, including fraud signals, transaction enrichment, credit scoring, recommendation systems, segmentation, and regulatory reporting.