Principal Data Scientist2013 - PRESENTGroupon
Technologies: R, Python, Spark, Hadoop, Shiny, Elasticsearch, Redis
- Built a user-to-product recommendation system that targeted customers with email and push notifications with high precision.
- Created a robust A/B testing framework that locates the best variant on any user-defined attribution method.
- Developed an automated bidding system for display, and SEM marketing.
- Built a product-to-product-based recommendation system to create greater engagement, and drive purchases.
- Designed a model to predict the Customer Lifetime Value (540 days) after any predefined action taken by a user.
- Developed a real-time, multi-touch attribution model to help identify the efficiency of the campaigns run by paid marketing channels.
- Created a framework to calculate the marketing channel incrementality, that helps a business identify which campaigns generated incremental orders, and which took credit from the attribution methodology.
- Designed a customer segmentation methodology to cluster users based on long-term profitability rather than immediate transactions.
R&D Analyst2012 - 2013Global Analytics
Technologies: R, Python, SQL
- Built an automated model building platform. Given any dataset, the module will attempt to build an ensemble model for the best possible result.
- Developed a user-defined variable generation module, which transforms the dataset based on the configuration file provided by the user.
- Built a module for missing value imputation, which helps a dataset utilize all possible information provided, and replace missing data with a proper guess-using simulation.
- Ran multiple ad-hoc analyses, which helped shape the business based on my recommendations.