Data Scientist And Engineer, Well-Being Team
Menlo Park, Ca, Us
Instagram’s Well-Being team improves the integrity and quality of content on the app. I’m the lead data engineer on the Discovery Safety sub-team, where I reduce harmful content in the search engine. Some of my work includes:- Owning and improving the team’s A/B experimentation platform and processes, including metric ideation and implementation, evaluating and tracking results, and negotiating tradeoffs between integrity and engagement- Securing $1.7 million in funding for, building, and calibrating a human labeling pipeline to more accurately measure the prevalence of harmful content across search surfaces- Implementing metrics to measure the prevalence of self-harm and eating disorder content in search results, which unblocked interventions that reduced harmful content by over 10% across Instagram search- Performing wide-reaching analyses on how teenagers engage with Instagram search differently from adults, which influenced our and the Teen Safety teams’ priorities for where to most aggressively reduce harmful content.Previously, I was the lead engineer on Facebook’s News Team for the News Ecosystem Quality (NEQ) score, which demotes low quality news content across the platform. Managed all aspects of the score, including:- Leading model implementation and upgrades across multiple news surfaces including Facebook’s News Tab, Watch video feed, and political content during the 2022 Brazilian presidential election- Improving the NEQ model’s accuracy on video content by over 40%- Designing a project for and managing an intern to improve monitoring of the model performance and infrastructure of NEQ
