• Worked with HealthDocs Ireland to build a Machine Learning pipeline that enables the firm to automatically extract relevant data from scanned PDFs and JPEGs using Object Character recognition (OCR)• Compared OCR algorithms across Azure, GCP and AWS to recommend the base algorithm which can be built upon

• Initiated a Machine learning project aimed at reducing vehicle downtime at dealerships, thereby increasing customer satisfaction• Lessened the run time of an Alteryx pipeline from 24 hrs to 2 hrs by programming it in Python• Collaborated on Virtual technician Assistant aimed at quicker diagnosis of a vehicle repair and decreasing vehicle downtime using NLP. The model had a top 3 accuracy of 65% during the pilot stage

• Constructed an R pipeline enabling a Retail client to forecast at scale by comparing over 8 algorithms and picking the best technique. The pipeline returned a MAPE of 10%• Coded a Modular Classification pipeline in python which allowed the client to not only read and prepare the data but also build, compare and pick the best among 4 well-tuned classification models• Continuously mentored Analysts and Trainee Analysts on various projects• Structuring and rolling out periodic reports

• Applied Random Forest to provide a client with a set of stores that have the potential to be converted to premium experience stores with a precision over 0.95• Built a Market Mix Model for a QSR client equipping them with the right insights to increase the spend on Digital Channels to 45% which would lead to a 4.5% increase in transactions with the same budget• Provided actionable insights at a store level for a leading QSR chain in order to turn negative EBITDA stores to profitable units.• Menu rationalizing and pricing activitiy for the client to improve operational health and gross margins.

• Forecasting demand of magazines and their allotment at a store level based off of multiple parameters along with optimizing the number of unsold copies.• Using statistical analysis to provide insights to the marketing team in order to facilitate better negotiations with the wholesalers.• Performing issue and cover specific analysis to push copies in high potential stores in order to generate maximum sales.• Responsible for testing and implementation of best display promotion at major retail stores for the top Time Inc and competitor titles like Time, Life, People, Sports Illustrated, National geographic etc.• Conducted extensive market research along with competitor analysis to identify latent sales potential at various retail locations across USA and Canada.• Iterated through multiple regression based models for predicting sales for Instyle magazine.• Automated manual process for various teams across the floor, hence saving hours of manual work throughout the organization.• Standardized and streamlined multiple databases, thereby bringing in ease of access.

Collecting the required Data by being on field and compiling it in a spreadsheet