Own and extend production batch and streaming pipelines that supply data and features for machine learning model training and online inference.
Build event-time streaming systems that handle late-arriving and out-of-order events, while ensuring point-in-time correctness and preventing data leakage and training-serving skew.
Develop and maintain production software in Python and SQL, including data transformations, integrations, and services supporting the ML data platform.
Build reliable processes for data quality monitoring, freshness, schema evolution, feature versioning, backfills, replays, and recovery when upstream data feeds fail.
Design and operate systems that deliver near-real-time features to ML models at low latency, while managing pipeline performance, reliability, and operating costs.
Col