Python · TensorFlow · Kubernetes · Docker · Java
ABOUT THE ROLE
Own the recommendation and matching stack end to end at an early-stage AI company in the advertising and consumer modeling space. You will build the models, real-time serving systems, and data platform that power personalized ad selection.
WHAT YOU'LL DO
- Build retrieval, ranking, and re-ranking systems for matching ads with recipients.
- Develop embedding and vector-search infrastructure for real-time ad selection.
- Optimize ranking across click-through rate, advertiser return, publisher revenue, and subscriber relevance.
- Own the feature store and serving path, including offline training, online inference, and pipeline reliability.
- Apply language models and foundation models to subscriber tagging, content understanding, and user modeling.
- Lead projects from prototype to production, including online learning, feedback loops, ETL/ELT pipelines, and warehouse transformations.
- Create dashboards and automated scorecards for performance and key recommendation metrics.
- Partner with company and engineering leadership on priorities, and help grow the ML function over time.
WHAT WE'RE LOOKING FOR
- At least 3 years of experience with recommendation, ranking, or advertising systems, and typically 5 to 16 years of overall experience.
- Production experience building a recommendation or matching system across candidate generation, ranking, embedding-based retrieval, serving, and evaluation.
- Experience with ad ranking, click-through or conversion prediction, or ad matching.
- Hands-on experience with embeddings and real-time vect