Emanuele Bardelli from OLX Group presented on their transition from batch to near-real-time data processing using Apache Spark. OLX operates classifieds platforms across 40+ countries and processes over 4 billion events per day. They moved from processing data once per day to processing data in small chunks within 10 minutes of events occurring. This allows them to send targeted messages to users in near-real-time. Spark is used to continuously load, transform, and store features from raw data files incrementally as new files are received. This enables more creative and flexible user communication through their lifecycle management services.