The posting, in OpenAI's own words
archived Oct 21, 2026About the role
We’re looking for a product data scientist to help OpenAI create more value and drive sustainable growth for young people, families, and learners. You’ll drive product data science across under-18 audiences, emerging opportunities for children under 13, family experiences, and learning and education more broadly. You’ll help decide where we should invest, what we should build, and how we’ll know it is working. This means understanding distinct audience needs, identifying promising use cases, and turning evidence into a clear product strategy. Your work will connect discovery and early product development with adoption, retention, and sustained user value. You’ll partner across Chat Product, Growth, Foundations, SafeSys, and other teams, working closely with product managers, engineers, designers, and researchers. The role focuses and owns product value and growth while dedicated teams own policy and safety.
In this role, you will:
Shape strategy and prioritize opportunities. Size unmet needs and growth opportunities across audiences, use cases, and product experiences. Help teams choose where to invest and sequence work based on expected user value, reach, evidence, and effort. Define what value and success mean. Develop an audience-specific measurement framework spanning activation, task success, learning progress, satisfaction, and durable retention. Connect these measures to product goals and assess whether repeat use reflects meaningful benefit. Understand users and growth drivers. Analyze journeys, funnels, and cohorts to uncover where people find value, where they struggle, and what helps them return. Combine behavioral data with surveys and user research to explain differences across ages, family roles, and learning contexts. Measure what changes actually cause. Design experiments and holdouts to evaluate product and model changes. Apply causal methods when randomization is impractical, make assumptions explicit, and account for seasonality, audience composition, and effects that differ across segments. Turn findings into product decisions. Work with partners from initial hypotheses through launch and iteration. Recommend what to build, change, expand, or stop, explaining expected impact, uncertainty, and tradeoffs. Follow through to determine whether decisions delivered the intended results. Build reliable measurement. Partner with engineering on instrumentation, dataset quality, and consistent audience definitions. Create reusable analyses and dashboards that make progress visible and let teams answer recurring questions independently. Create alignment across teams. Bring a coherent view of audience value to work spanning multiple roadmaps. Help partners agree on priorities, success criteria, and the next questions to resolve, even when ownership is distributed and evidence is incomplete.