Lead and launched text categorization platform processing 40k B2B news sources and 1.4M paired / classified articles (monthly) using nlp, text analytics, deep learning, and streaming pipelines. Increased precision and recall at the same time to 90% and 84% respectively. Team included 1 product owner, 2 SCRUM teams (12 data scientists / engineers), and 18 months of R&D. Foundation to entering B2B text IR category and new company signal monetization (as-a-service).