Head of Data Science, Intelligence Ventures

This role requires the ability to work lawfully in the U.S. without employment-based immigration sponsorship, now or in the future. The Head of Data Science builds the models that turn a new AI-native behavioral intelligence platform into products customers buy, and proves to those customers that the products work. At the core of the platform, part of Spectrum Intelligence Ventures, is a proprietary transformer-based model that learns how households behave from network-scale signal and encodes that behavior as embeddings. Those embeddings are the foundation for audiences, insights, measurement, and data licenses sold to brands, agencies, and ad tech, media, and AI platforms. Over the next 18 to 24 months, this leader's job is to get that portfolio to market and make it stand up to buyer scrutiny: build the predictive models and products on top of the embedding layer, set the evidence standard buyers will test them against, and work alongside product and sales to scope pilots, answer technical buyers, and convert evaluations into contracts. In parallel, this leader owns the direction of the embedding model itself, working with the Head of Technology and external development partners, and is accountable for showing that it delivers performance conventional approaches cannot. The Head of Data Science builds and leads the data science team, sits on the platform leadership team, and reports directly to the Head of Intelligence Ventures. HOW THE HEAD OF DATA SCIENCE MAKES AN IMPACT * Build and run the predictive models behind the platform's audience and data products, including in-market, life-stage, purchase propensity, category interest, lifestyle affinity, and behavioral velocity models scored daily at the household level across more than two dozen verticals. Own segment definitions, coverage, refresh cadence, and measured accuracy for every audience the platform sells, whether through Spectrum Reach, DSP data marketplaces, or partner platforms. * Own the platform's Feature Store as a commercial product: named, typed, versioned, and documented features that serve audiences, insights, measurement, and partner models. Lead the outcome anchoring program that trains and calibrates models against licensed third-party outcome data (vehicle registrations, mortgage originations, card and SKU-level retail spend, travel bookings, and similar sources) so that propensity scores mean what they say, and automate feature creation as new signal sources come online. * Lead the modeling behind custom and partner-specific audiences, including seed-based lookalike expansion, profile-based audience construction, and site- or CRM-seeded models that identify households entering a brand's purchase path before they reach the brand. Establish repeatable, clean-room-based workflows with transparent match rates, expansion multiples, and performance against the buyer's own baseline. * Design and defend the platform's measurement methodology, including exposed-versus-control campaign lift using matched behavioral controls, test-and-holdout incrementality, and closed-loop validation against real-world outcomes. Define the journey, demand, and category metrics that power insights subscriptions and aggregate demand indices, built to withstand scrutiny from brand marketing science teams, agency analytics groups, and measurement partners. * Lead the data science behind the platform's identity and viewing products: the adjudication and confidence calibration methods that verify partner household graphs against network ground truth, and the household viewing-state, trajectory, and reachability models that underpin viewing intelligence for programmers, agencies, and measurement companies. * Direct the roadmap for the platform's proprietary transformer-based behavioral embedding model, a multi-entity architecture that encodes both a stable household profile and short-term intent, working with the Head of Technology and external