About this role
Summary: Design and operate real-time personalized ad recommendation systems, including retrieval, ranking, and vector search. Lead data warehouse transformation and feature store management while collaborating with executive leadership to define technical priorities and build the ML team.
Responsibilities
- Design and run the full matching system that serves real-time personalized ad recommendations, including retrieval, ranking, and vector search.
- Build ranking methods that balance several goals at once: click performance, advertiser returns, publisher revenue, and subscriber relevance.
- Keep the feature store accurate, remove differences between training and live serving, and own the health of the data pipelines.
- Lead the data warehouse transformation layer (dbt), subscriber tagging with foundation models, and automated scorecards and dashboards.
- Work directly with the CEO and CTO to set technical priorities, with a path toward building and leading the ML team.
Requirements
- 5–16 years of hands-on production experience in machine learning, recommendation systems, or data infrastructure.
- Deep expertise in vector search, ranking algorithms, feature stores, SQL/dbt pipelines, and real-time model serving.
- Able to work hybrid in the New York City area.
- US citizen or Green Card holder only. No visa sponsorship available.
Skills Required
- Recommendation & ranking systems
- Vector search & embeddings
- Feature stores & online inference
- SQL, Postgres, dbt
- Snowflake / BigQuery / ClickHouse
- ETL/ELT pipelines
- LLMs & user behavior modeling
- Multi-objective optimization