Employ a variety of data science techniques including natural language processing, machine learning, deep learning, statistical analysis and data visualization using Python and R to solve key business problems for internal and external clients• Design and prototype data science experiments for business development and client efforts• Lead data science projects from idea inception to product deliveryKey achievements include:• Created a natural language processing pipeline using DistilRoBERTa in Pytorch for supervised text classification to automate web-based interviews scoring. The model significantly outperformed the previous approach in terms of man-hours and costs reduction• Spearheaded a topic modeling effort to classify survey verbatims. The research used an unsupervised approach at first to gather overall topics. These topics and their derivatives were used as anchors for a semi-supervised model• Analyzed an executive's tweets for public relations and image management. Built sentiment analysis, benchmark models for comparables executives and a topic model using replies from Twitter • Helped build a tree-based model to predict employee turnover for a large healthcare client. The model generated a five-fold improvement in lift• Designed and created a multi-step research to understand mobility within a large organization. Implemented an employee mobility predictive model using One-Vs-Rest classifier with a tree-based approach. Built a cluster analysis model to create personas with the goal of understanding high and low mobility clusters• Created a multilevel regression and poststratification model using national survey data to estimate state-level opinions for a public sector client. The model estimates were used to create a data visualization dashboard using Tableau for public officials.

• Analyzed customer behavior and built a sales prediction classifier – increased sales by 5%• Built a call center representative scheduler optimizer by analyzing incoming calls and operations data. Optimized operations for a campaign and reduced costs by 9%. Deployed the production model on AWS. • Ad-hoc analysis for clients• Client-facing role - created pitchbooks and presented to clients on capabilities