Led efforts to collect, analyze, and leverage data for nearly every product across the company.Data engineering - Integrated with partners and clients to gather data to support products and drive user engagement. For example, integrated with food service providers to collect information about recipes, restaurants, and groceries to be used in a variety of products. Built out integrations with tools such as email service providers to allow for engaging email campaigns. Drove the design and implementation of a flexible survey system to gather data from users. With so much variegated data flowing through our system, I designed and built systems for ETL and experimentation so that non-engineers can understand and visualize usage trends while respecting the privacy of users’ health data.Machine learning - Leveraged novel machine learning and natural language processing techniques to make structured and unstructured predictions. For example, made predictions about food flavors, nutrition, etc. robust to varying levels of data quality. Modeled user health and nutrition by marrying techniques from the nutrition literature with our proprietary data. Paired our understanding of food and nutrition to build out scalable search and recommendation systems to improve user health.

Mined and analyzed data from a variety of sources. Applied large-scale machine learning techniques to solve prediction problems which power user-facing products. Developed tools and infrastructure to make data manipulation and analysis easier company-wide.As a technical lead for the User Modeling team, I led efforts to improve and consolidate machine learning algorithms for predicting user attributes. Worked closely with both front-end and infrastructure teams to develop impactful predictions.I was also a founding member of the Entities team which manages structured data on Facebook. Helped develop techniques for entity extraction, deduplication, disambiguation, etc. Worked on a number of classification problems such as authenticity. Focused especially on building databases for music and places and was involved in the launch of those integrations on Facebook.As a member of the Core Data Science team, I used novel statistical machine learning to analyze the Facebook user base. Projects included inferring user ethnicity, analyzing intra- and inter-ethnic interactions, and predicting user engagement. Blog posts are available for some analyses below.