• Build data pipelines: Develop robust batch and streaming pipelines (Kafka/Kinesis) to ingest, transform, and enrich large-scale event data (impressions, clicks, conversions, costs, identity signals). • Create data aggregates & marts: Design and maintain curated aggregates and dimensional models for multiple consumers—prediction models, agents, BI dashboards, and measurement workflows. • Data modeling & contracts: Define schemas, data contracts, and versioning strategies to keep downstream systems stable as sources evolve. • Data quality & reliability: Implement validation, anomaly detection, backfills, and reconciliation to ensure completeness, correctness, and timeliness (SLAs/SLOs). • Performance & cost optimization: Optimize compute/storage for scale (partitioning, file sizing, incremental processing, indexing), balancing latency, throughput, and cost. • Orchestration & automation: Build repeatable workflows with scheduling/orchestration (e.g., Airflow, Dagster, Step Functions) and CI/CD for data pipelines. • Observability for data systems: Instrument pipelines with metrics, logs, lineage, and alerting to accelerate detection and root-cause analysis of data issues. • Security & governance: Apply least-privilege access, PII-aware handling, and governance controls aligned with enterprise standards.