Mansi Borkar

Data Modeler II

Experian•  July 2021 - Present

Streamlined data pre-processing pipeline by building a scalable data pipeline using HiveQL, Python, and AWS; reduced data processing time by 50% and improved algorithm accuracy by 10% • Constructed and deployed Machine Learning models on consumer credit cards behavior across 10+ portfolios; delivered a 15% increase in revenue and a 20% increase in approval rate while mitigating risks • Implemented feature selection and Bayesian optimization hyper parameter tuning techniques to enhance model performance and improve data-driven decision-making capability; key driver in securing a major new client, increasing the company's revenue by 10% • Spearheaded the development of Statistical Modeling with cross-functional teams; optimized product strategy to increase response rates by 12%, reduce acquisition costs by 8%, and drive a 20% increase in marketing ROI • Created and presented customized reports to clients, outlining the impact of automated lending decisions on business operations and performance metrics, including a reduction in loan processing time by 50%

Data modeling analyst

Experian•  January 2021 - May 2021

• Extracted, cleaned, and analyzed data for 10+ clients in the finance industry using scripting language Python, R, and HiveQL; identified trends /patterns/correlation and resolved data inconsistencies, resulting in a 20% increase in data accuracy • Streamlined Prescreen campaigns by identifying creditworthy consumers using internal data analysis; increased response rate by 12% and earned a 20% boost in marketing ROI • Orchestrated the design and creation of interactive dashboards and Data Visualization through Tableau, proposed business solutions that enabled stakeholders to make informed decisions and increase customer satisfaction by 15%

Graduate Assistant

California State University Fullerton•  October 2019 - December 2020

• Led an initiative to build a sophisticated Machine Learning Algorithms that predicts temperature changes in 69 zip codes, and calculates optimal solar panel count; deployed model using Python • Engineered and developed a Natural Language Processing (NLP) model using Python to classify 6,000+ articles as fake or real with 90% accuracy, which enables to make more informed decisions on content strategy and distribution • Spearheaded the development of a Decision Tree model that predicts customer churn with 90% accuracy using R programming, Data Mining and Pruning techniques

Data Analyst

State Street Corporation•  December 2016 - June 2019

• Developed and executed test plans for ETL processes, detecting and fixing 75+ issues; improved data quality and saved 20+ hours of manual testing each week • Orchestrated cross-functional teams to deploy production releases in record time by streamlining communication channels, reducing downtime by 15% and increasing overall efficiency by 8% • Trained and coached 4 junior data analysts on advanced SQL query writing, data extraction, Tableau data visualization, and dashboard design; elevated team efficiency by 30% and reduced report turnaround time by 2 days

University of Mumbai

Information Technology, BE•  July 2012 - June 2016

Advanced database management system, Big Data Analytics, Distributed database