University of California, Berkeley

MOOCs

Senior Data Scientist at Opendoor

San Francisco, California, United States

- Leading a company-wide effort to deploy an ML platform for DS initiatives.- End-to-end designed, researched, and productionized a recommendation engine for our institutional buyers. Customers can find 80% of true leads in 1/3 of the time.- Built a rental valuation model to produce state-of-the-art rental value estimates (>10% error reduction compared to next-best).

Opendoor

Senior Data Scientist

- Proposed, scoped, researched, and productionized a (Scala/Spark) classifier end-to-end, and contributed to other regressors and classifiers- Developed a novel approach to WiFi-based localization with 20% reduced error compared to standard practice- Collaborated cross-functionally to build client-facing dashboards and visualizations, providing an intuitive understanding of statistical insights- Lead research projects, compiled results, and presented them across teams

WeWork

Data Scientist

- Improved heuristic matching algorithm and supervised learning pipeline resulting in 10k’s of additional matches for core dataset with 60%+ reduction in error- Conducted product research to inform key stakeholders and distilled for various audiences (e.g. data, product, company-wide presentations)- Wrote production Scala/Spark code operating on a dataset with trillions of rows- Educated product team on SQL and data fundamentals

Euclid (acquired by WeWork)

- Built ML models to predict job candidates’ aptitudes with >2x the performance of traditional interview assessments.- Wrote scaleable, production code for ML models- Partially automated model building, cutting development time by >50%, and partially automated data collection, reducing time investment by >75% and cost by >30%- Oversaw the entire experimental pipeline - design, collection of data, analysis, and model building- Collaborated with the product team to develop roadmaps, features, and dashboards

Scoutible

Organized and restructured data, ran statistical analyses, visualized data, focusing on harm reduction in opiate users

UCSF

Research Assistant

Designed and programmed experiments; cleaned, formatted, and statistically analyzed dataVisualized, graphed, and presented resultsDeveloped a model of visual aesthetic preferences based on evolutionary utility

Jan Flatley-Feldman