Sevatec
Étude de cas client
Sevatec is a high-technology services firm that engages trusted talent and leverages technology innovation to overcome the federal government’s most pressing challenges. Sevatec supports a number of mission-critical applications across the federal government, to include Homeland Security, Department of Defense, Department of Transportation, Department of State, and other federal and civilian agencies.
Cas d’utilisation verticale
- End-to-end analytics to deliver machine learning applications
Cas d’utilisation technique
- Ingestion de données et ETL
- Analyses SQL
- Machine learning
Témoignage clientLes défis
- Difficult to ingest and prepare data across 30+ disparate systems
- Support a 2000+ users who are siloed and have different skill sets (BI users, statisticians, data engineers, data scientists, business)
- Lacked the ability to prepare a single view into the data for data science
- Inability to scale their data science efforts as they were using RStudio on a single node
- Used a variety of disjointed tools to perform large data extractions which created substantial DevOps complexity
- System was being heavily taxed by very IO intensive queries, impacting their ability to meet SLAs and the demands of the rest of their user community
La solution
Sevatec leverages the Databricks Unified Analytics Platform to simplify data analytics for Federal agencies. They’ve greatly simplified data engineering through a fully-managed, cloud platform and has accelerated data science innovation by fostering a culture of collaboration and transparency.
- Democratize access to data across their various data sources through APIs and data source connectors. Reduced access an ingest times from hours to minutes.
- Able to build build machine learning models at scale against the entire data set
- Simplified infrastructure management and eliminated unnecessary DevOps work through automated and secured cluster management
- Interactive workspace allows various users to collaborate on the data and run data science experiments that lead to innovative machine learning models
With Databricks, we get the secured, easy access to the data without worrying about the engineering behind it.