Accelerating autonomous vehicle development
Hours to ingest data, compared to 2–3 weeks before
Shaved off client development cycles
Reduction in cost per file processed
“Without Databricks, we wouldn’t be able to process the sheer volume of data needed to help our customers develop better autonomous vehicle capabilities and drive their product roadmap. We now have a repeatable process to use across our growing client base.”
— Patrick McAuliffe, Lead Engineer, Incite
Incite’s Rapid OEM Automotive Data (ROAD) platform enables the world’s largest automakers to build safer autonomous vehicles. By processing the massive amounts of unstructured data generated by test fleets, ROAD enables automakers to deploy autonomous driving features faster and with greater confidence. Before implementing Databricks and building the ROAD platform, Incite ran into frequent performance issues, which made it difficult to deliver fleet behavior metrics that clients could consume easily. Meanwhile, customers had to spend hours manually searching individual files for the data they needed, significantly slowing down product development.
End users get new data in hours, not weeks
To get the best qualities of data warehouses and data lakes, Incite integrated the Databricks Lakehouse Platform into its ROAD platform. Databricks processes hundreds of terabytes of data points and scales to support teams working on capabilities such as self-driving on highways, preventing parking lot collisions and navigating neighborhood streets. Incite can now ingest terabytes of data and make it visible in end-user dashboards in less than 12 hours, compared to two to three weeks previously.
With Databricks SQL, Incite’s clients perform structured and ad hoc analysis to explore host vehicle and feature event behavior across many dimensions of vehicles, driving scenarios, vehicle software versions and driving conditions. To correct bad client data, the company uses Databricks Delta Lake rather than performing time-consuming rollbacks. Incite is also working with Databricks to develop stronger geospatial capabilities that will help detect anomalies in clients’ software and prevent car crashes.
Software development cycles go from weeks to days
With Databricks at the heart of its ROAD platform, Incite is helping its clients in the automotive industry meet their ambitious goals of delivering revolutionary autonomous vehicle technology. The company’s clients have shortened their software development cycles from weeks to days. Even with significant increases in data volumes and the number of files processed, Incite has reduced its costs per file processed by 50%, passing these savings on to clients. Clients, in turn, have shown their satisfaction by expanding their use of the ROAD platform — which will mean safer roads for everyone in the long term.