A research team at BrightVerge Labs ingests about 36 TB of raw text each week from PDF files, chat transcripts, and application logs that arrive through Cloud Storage and PubSub. They need a highly ...