Uber Maps Org - Focussed on improving Search

- Developed a scraping system that continuously pushed fresh job data from other sites onto Pared's jobseeking app system which resulted in a 40% increase in the number of total viewable jobs.- Automated Pared's gig worker suspension policy. Developed a system that issued suspensions, deactivations, and reactivations, which saved the support team 8-10 hours every week and increased the available pool of gig workers by 7%.- Implemented full visibility of a customers journey from the top of funnel to activation. Unified analytics across web and mobile, migrated Pared's web app to a new domain, and developed SQL views for customer journeys.-Enhanced the friend suggestion feature in pared's mobile app by unifying several separate data sources. This work resulted in a 7x increase in the average number of suggestions made and doubled the average weekly number of connections made.

- Designed A/B tests for product and marketing experiments. Also tracked, explored, and defined success metrics for new features. Some learnings include price insensitivity for promotional deals and customer rating apathy.- Supported Engineering by fixing bugs, performing data migrations, and writing production queries for Pared's payment system.

- Created models in Python using numpy, pandas, and well known machine learning libraries (scikit-learn, xgboost, keras) to predict the onset of different medical conditions. Notable models achieved an AUROC of 0.8-0.85 for the prediction of sepsis and actute kidney injuries.- Worked with operations to help manage customer relationships. Notable projects: Implemented a live monitoring system for each hospital and developed medical tests and summary reports for quality assurance of a diagnostic product.