Lead Data Scientist
CurrentLed the development and deployment of machine learning models for natural language processing, recommendation and propensity-to-subscribe scoring.- Vertex AI Integration: Developed and deployed custom and AutoML pipelines using Google’s Vertex AI, overhauling the team’s machine learning workflow. Deployed a user conversion propensity model using Vertex that improved model development and deployment velocity by 6x over previous processes.- AI Narration Service: Launched The Atlantic’s first AI-powered article narration service on-site, scaling our ability to narrate articles 20x faster than existing non-AI narration partners. This met a longstanding user request for more audio content on our site.- NLP and Topic Modeling: Designed and implemented a Latent Dirichlet Allocation (LDA) topic model, classifying The Atlantic's content into machine-discovered themes. This model supports content strategy and recommendation.- Article Recommendation Engine: Architected an in-house recommendation engine using content-based and collaborative filtering techniques to replace existing logic on all on-site articles, resulting in a 41% increase in the recommender’s CTR. - Metadata Enhancement: Advanced content tagging processes with sentiment analysis and named entity recognition, significantly improving metadata quality and facilitating better content analytics capabilities.- Site Traffic, Programmatic Ads and SEO Analysis & Reporting: Conducted extensive analyses of site traffic, programmatic ads and SEO, and created comprehensive reports using BigQuery and Looker.
