Data Ingestion and Topic Modelling

Designed the staging data model and built data processing pipelines in Talend for ingesting data into a central data lake. Implemented topic modeling and sentiment analysis on transcripts of customer support call interactions that improved the time in feedback and decision analysis process by 40%. The entire data processing involved real-time model scoring with a daily data refresh of about 3-5TB. The client was one of the largest American telecommunications organizations.