Lead Data Scientist, AdTech

Role overview Build and operate machine learning models that guide which auto insurance offers shoppers see and how advertising spend is allocated. This hands-on individual contributor role covers problem framing, data discovery, model development, validation, deployment, and monitoring, with an initial focus on insurance and potential expansion to other market categories.

The work involves delayed conversion outcomes, sparse sessions, returning users, and changing market conditions. Models also influence which users are reached, so their decisions affect the data available for future learning. Success means producing dependable models that the business can use directly, including bringing automated marketplace bidding from proof of concept into scaled operation.

Responsibilities - Develop models that match shoppers with likely-to-convert carriers and inform media buying against return-on-ad-spend goals. - Incorporate conversion outcomes such as quotes, policies, and acquisition cost into production models. - Find and prepare relevant data, frame modeling problems with business partners, and validate solutions. - Work with an ML engineer to deploy models and monitor their performance in production. - Address late-arriving labels, repeat users, sparse features, market drift, and feedback effects from bidding decisions. - Take ownership of insurance marketplace media buying, then help extend the modeling approach as scope grows.

Requirements - At least five years of applied data science experience with measurable business results. - Experience building models that directly influence paid media spend, bidding, or ad serving; attribution and reporting alone are not sufficient. - Ability to explain and defend model design choices, including loss functions, architecture, and neural network fundamentals. - Experience with delayed labels, sparse data, drift, or the effect of decisions on future training data. - End-to-end ownership of models, from data discovery and development through deployment and monitoring. - Expert-level Python and SQL.

Nice to have - Experience spending real advertising budgets in live auctions or building large-scale ranking and recommendation systems. - Knowledge of auction feedback loops, selection bias, and exploration strategies; ability to explain tradeoffs to nontechnical partners. - Hands-on neural network work, including embeddings, or data science experience with AWS services such as ECS, S3, Athena, or SageMaker. - CI/CD deployment experience.

Benefits and work setup - Remote-first team; most colleagues work East Coast hours.