Rivian Databricks (2024)

Rivian’s Aspring Brick Climber Application

Databricks serves as our central Data Platform, and for a multitude of good reasons. By decoupling our storage architecture and keeping our cloud vendor securely in the “friend zone,” we’ve unlocked a tremendous amount of flexibility. We seamlessly employ open storage standards—like Parquet and Delta Lake—without a second thought. This architecture allows us to distribute compute from storage, meaning we can innovate and scale our analytical workloads almost anywhere, all while treating a simple object storage bucket as our single source of truth.

Deez Watts pipelines a continual JSON stream of raw Rivian telemetry data—capturing everything from battery state and ambient temperatures to motor torque and geocoordinates. This stream is ingested, refined, and served up as a high-performance Delta Live Table in Databricks, providing the ultimate foundation for rich, post-adventure data analysis.

By leveraging the “Medallion Architecture” (Bronze, Silver, Gold data layers), we ensure that our telemetry transitions from raw exhaust to finely tuned datasets, enabling advanced geospatial queries, trip replay features, and potential machine learning models to analyze our overland expeditions.