At WarnerMedia, I pioneered a project that utilized Nielsen TV data and consumer surveys to craft audience lookalike models through Random Forest regression, predicting demographic-based viewership on our linear TV platforms. We built an optimization model to enhance demographic-targeted impressions across all showtimes for advertisers. To cater to the Latin American market, we developed a self-service tool, empowering the sales team to autonomously define audience segments and determine ideal ad placements, which significantly assisted in sealing advertising contracts. As the lead product manager, I directed the engineering team, curated feedback loops with sales, and managed the development trajectory. This initiative transformed a lengthy manual procedure into a few-hour streamlined process via a custom React UI, integrated with machine learning models on AWS Fargate and Lambda, all housed in AWS Snowflake, setting a new standard for audience-targeted advertising.
