NBCUniversal partnered with EXL and Databricks to modernize its data platform, moving from a shared slot-based model to Databricks job-specific compute. The migration helped NBCUniversal reduce data infrastructure costs by 30%, onboard more than 300 analysts to Databricks SQL, and create a more flexible foundation for analytics, machine learning, and data engineering.
Key highlights
- 30% lower infrastructure costs: Dedicated compute lets pipelines scale independently instead of competing for shared reserved capacity.
- A unified analytics platform: Data engineering, SQL analytics, machine learning, and workflow orchestration now run together on Databricks.
- Phased migration with EXL: Custom tools supported discovery, SQL conversion, data transfer, orchestration migration, and automated validation.
- Safer cutover process: NBCUniversal ran workloads in parallel, compared results, and migrated in waves before retiring legacy workloads.
- Ready for changing demand: The new architecture gives teams more flexibility to scale around major content launches, awards shows, and live events without pre-reserving capacity for every workload.
👉 Read the full post here
