Created a model to re-rank search results based on the availability of restaurants to optimize mean reciprocal ranking of booked restaurants, resulting in 35% increase in MRRAnalyzed availability influence on user booking behavior, generated custom booking distributions and computed feature values from availability, distributions, and referring serviceCreated a Python ETL for cleaning and transforming data from S3 to cluster for model training, used PySpark and Python UDFs to efficiently compute custom features for millions of rowsBuilt features in Java with JUnit tests to add custom features and distributions to RankNet modelUsed Scikit-learn random forest to determine feature importance and visualize decision rules and Keras logistic regression to create a model from these features with performance comparable to productionParticipated in agile development process with a team, derived meaningful tasks from OKRs

Built serverless software using Amazon Web Services for a technology startup, provided backend for a leaderboard to rank participants on different factors at business networking eventsDeveloped a mathematical model for evaluating performance of a business networking productImplemented integration tests, using test driven development and behavior driven development to ensure functioning code before it reached productionParticipated in code reviews, used pair programming methods, presented and explained code

Built an ETL program using Ruby on Rails for an investment from the FSB Companies’ portfolio, organized data and presented notifications to describe data upload failuresCreated a front-end interface for the ETL using Bootstrap, displayed and sorted notifications by categoryLearned business perspectives and strategies for investmentExperienced a real-world, competitive environment for business and technology