Case Study: Auto Trader achieves 3x faster ML model deployment with Databricks

Auto Trader - Customer Case Study

AutoTrader, a UK online marketplace for new and used cars, needed to build a recommendation engine to predict car valuations based on age, mileage and condition. They struggled to scale a legacy stack to process massive volumes of data, faced handoff and collaboration issues between data scientists and engineers, and were slowed by DevOps complexity from EMR and Jupyter notebooks and limited experience productionizing Spark-based models.

Databricks provided a unified analytics platform with automated cluster management and a collaborative, multi-language notebook environment, letting teams more easily build, train and deploy machine learning models. The platform made data and analytics accessible across disciplines, improved collaboration between data science and engineering, streamlined model production and sped time-to-market for new models by 3x.

Auto Trader

Edward Kent

Principal Developer, Data Engineering