Ryan Tonucci

At Samsara, I architect streaming telemetry and GenAI data platforms processing 4B+ events daily from 2.5M+ assets. I built the retrieval platform for the fleet assistant, improving retrieval hit rate from 62% to 89% while reducing embedding compute by about 80%.

I’ve built ML-ready data systems across Samsara, Optum, and Databricks, including feature platforms, clinical NLP training pipelines, and exactly-once streaming ETL. At Optum, I scaled HIPAA-compliant claims and EHR pipelines for 100M+ members and standardized 300+ ML features across four teams.

I lead platform work from infrastructure through operations, using AWS, Azure, Spark, Kafka, Iceberg, Databricks, and Terraform. My work has reduced lakehouse spend by 35%, cut data incidents 60% year over year, and enabled faster model iteration.