Google Maps Mobile. Making it easier for you to find things.

• Attended two-month intensive training on software engineering best practices

• Migrated an internal model-fitting service to be compatible with a new version of Tensorflow

• Implemented network distillation to reduce large ensembles to more compact models while retaining the same predictive power

• Assist ninth graders twice weekly with basic computer science principles, including variables, loops, and lists

• Support a classroom teacher with limited computer science knowledge by enriching lesson plans and assignments

This talk goes into the technical details of transforming twenty years of Vogue cover photos into a numeric dataset that measures skin tone lightness, which, when visualized, showcases the striking lack of diversity on the cover of Vogue magazine.

• Redesigned a constrained clustering algorithm and improved a random forest model that are used to transform record-level data to person-level data

• Improved the time and space efficiency of a record-linkage tool that is run on decades’ worth of data from the NYPD

• Used RMarkdown to create a professional PDF filled with metrics and stats about a model’s prediction that is generated automatically after each run

• Created a visual debugger that helps identify areas for model improvement

• Implemented extensive test cases and documentation to prepare for making internal software open source