Mitigating financial fraud at Cash App with ML solutions

• Technical lead for data science team, building flu / COVID forecasts with epidemiological and ML methods. Worked closely with product teams to scope and roadmap deliverables• Designed and deployed three data products tracking COVID-19 outbreaks using time series forecasting models and anomaly detection on geospatial sensor data collected across the US• Automated forecasting analytics to highlight the value of Kinsa data to inventory demand forecasting. Presented the business value for B2B clients in numerous sales meetings• Led redesign of county-level, real-time illness incidence data product generated from user temperature readings. Implementing quality control, smoothing, and interpolation methods

• Created a map-based web app for consumers to view their exposure risk to airborne pollutants using monitoring data collected from Google Street View cars• Engineered over 20 geospatial features from a wide variety of open-data sources, including OpenStreetMaps, US Census, city zoning GIS, and weather data• Predicted pollutant exposure with random forests with R2 of 0.95. Provided alternate housing recommendations based on model predictions and web scraping Zillow rental prices

• Processed and analyzed high-frequency greenhouse gas monitoring data collected across 9 sites to assess the viability of wetland restoration under California’s Cap-and-Trade program. Published results in 3 first-authored peer-reviewed publications• Applied random forest regression and information theory-based causal inference to understand drivers of increasing wetland greenhouse gas emissions during the California drought - https://bit.ly/2NEVBW4• Informed future wetland restoration site selection by improving our understanding of ecosystem methane drivers using wavelet time-series decomposition and mutual information statistics - https://bit.ly/3sczK7z