This document summarizes a presentation about Pollfish, a mobile survey platform. It discusses how Pollfish collects data from over 170 million mobile devices worldwide through surveys and device/user information. It handles big data by storing raw data in HDFS and using Cassandra for time-series and user profile data while PostgreSQL stores aggregated data. Spark is used for general cluster computing. Machine learning models use survey data to classify users into personas to personalize surveys. The system architecture includes HDFS, Cassandra, PostgreSQL, Azkaban for workflows, and Spark.