At Sorenson Communications (Hand Talk), I worked on the Data Platform team building a lakehouse on S3 and Apache Iceberg for an AI sign-language translation program. I implemented Amazon DataZone for the first onboarded source and designed the platform’s first data contract using ODCS.
At Grupo Trapézio, I built a PostgreSQL Data Warehouse from scratch, bringing together three disconnected ERPs in a Medallion architecture. My ETL work cut a critical dashboard from 40 minutes to 3 and reduced BI costs from over R$20,000 per month to under R$5 per month.
At Ademicon, I built and owned around 50 production pipelines processing millions of records daily, including continuous CDC with AWS DMS, PySpark, and Apache Iceberg. Earlier, at BRF S.A., I migrated more than 100 legacy SQL Server scripts to Azure Databricks and put five machine learning models into monitored production.
