• Led the development of a critical model to mitigate the loss of a fundamental data source to maintain customer satisfaction, with a significant impact on customer retention.• Developed an outlier detection product from incubation to delivery, to handle significant outliers in training download and revenue estimation models, that resulted in 80% reduction in customer tickets.• Collaborated with cross-functional teams, including Data modeling, Data analysts, PMs, and TPMs, to ensure timely delivery of requirements into production.

• Led development a customer segmentation model to prioritize prospects which resulted in 20% more efficiency of the sales team and growth in demand.• Led the developed a demand prediction product for a medical supplier company that resulted in 10% efficiency in the supply chain.

• Technical lead on a global cross-functional product across sales, servicing, and analytics orgs to design and develop PayPal's Merchant Health Score, which resulted in 2K+ internal user adoption.• Created a contextual churn and decline (C&D) model that identified the cause of churning, leading to a 33% reduction in CSM response time and an 8% increase in handling C&D cases.• Developed an award-winning deep-dive analysis product that democratized access to data for a team of 20+ financial analysts reporting to the CFO of North America, reducing the analysis time by more than 400%.

• Created a targeted review feature extraction tool that extracts insights on user specified aspects of a product• Deployed an interactive web-app to visualize the findings on AWS using Python, Flask, HTML and CSS

Working in an interdisciplinary research group leveraging techniques from various backgrounds to add value to real-world research problems. Select projects: Real-time Waze feed clustering• Created a real-time clustering method for raw crowdsourced Waze traffic data• Reduced the raw data feed by 60% through clustering• Enhanced the feed by providing the traffic queue length, duration, and reliability of the reports• Introduced customized measures for cluster validation and parameter tuningWaze data evaluation• Created quality metric for Waze data in comparison to existing data sources• Analyzed the value of Waze data for the Iowa DOT, documenting strengths and weaknesses• Secured additional funding to continue Waze data analytics Traffic incident prediction from social events• Scraped Chicago’s social events from the web and traffic incidents from Waze on Twitter’s REST API• Developed a traffic incident prediction model based on social events in Chicago• Conducted Twitter text analysis in Iowa for further traffic data collection and model enrichment• Supervised a team of two undergraduate research assistants in this projectDriver stress analysis • Conducted physiological data analysis to study drivers’ stress in naturalistic conditions• Modelled driver stress using heartrate and the electrodermal activity (EDA) • Analyzed environmental factors in driving that contribute to stressReal-time Traffic Incident Detection using speed data (anomaly detection)• Used Sparse Principal Component Analysis (SPCA) as an innovative detection approach • Working on expansion to multivariate data (tensor)

In a Marketing Strategy development Project for The National Iranian Drilling Company (NIDC): - Analyzed opportunities and threats according to the national 20-year vision. (SWOT analysis) - Implemented an AHP (Analytical Hierarchy Process) method to prioritize the target markets. - Delivered the official project report and documentation.