Lead Data Scientist
CurrentTook a NLP proof of concept in a SageMaker notebook, and spearheadedeffort to turn the logic into an event driven ML pipeline using Lambdas,custom asynchronous Sagemaker Endpoints, RDS, & QuickSight. Used inproduction on a daily basis. Final product- Identifies pertinent topics by extracting keyphrases in anunsupervised manner using pretrained language models andparts-of-speech-ML pipeline derived keyphrases from ~350K documents andsummarizes results on QuickSight dashboard. Pipeline automaticallyscales to handle larger than average data pulls.- Dashboard used to identify current & emerging topics for businessowner of dataset-Single click deployment of pipeline via CloudFormation and Jenkins