Data Scientist, Personalization & Machine Learning

Job description

About the Team We work across product, growth, and monetization to help people get more value from OpenAI’s products. We identify opportunities for machine learning, build and test early solutions, and partner with engineering and product to bring successful approaches into production. About the Role We’re looking for a data scientist who can enter an unfamiliar product area, find a useful ML application, and turn it into a working product intervention. You will own the work from problem definition and data preparation through modeling, experimentation, and launch evaluation. Projects will include personalization, next-best-action recommendations, lifecycle management, promotion targeting, and models of retention and customer lifetime value. You will help identify opportunities, build proofs of concept, evaluate, and iterate. In this role, you will: Identify opportunities with product and business partners, define the decision a model will inform, and prioritize work by user value, feasibility, and expected impact. Build datasets, features, reproducible training workflows, and models. You’ll use a combination of tools appropriate to the task at hand. Translate predictions into product actions. Design experiments and use causal reasoning to establish whether those actions improve user and business outcomes. Partner with engineers on integration, serving or batch scoring, monitoring, and rollout. Stay involved through launch and iteration. Explain results and limitations clearly, document decisions, and build reusable methods that other teams can adopt. You might thrive in this role if you: Have built and evaluated applied ML systems that influenced a product or business decision, with clear ownership of the modeling work. Write effective Python and SQL and can investigate imperfect data, debug models, and make pragmatic implementation choices. Build and schedule production scoring pipelines using dbt, airflow, and flyte. Understand supervised learning, evaluation, experimentation, and the difference between predicting an outcome and changing it. Can scope ambiguous problems, work across disciplines, and explain technical tradeoffs to colleagues with different backgrounds. Are motivated by helping people use AI effectively and by learning unfamiliar product domains. Are excited to take on challenging, impactful ML problems across multiple areas. Have experien…