- Own the data ingestion layer for device telemetry, transaction events, and KYC signals using streaming (Pub/Sub, Beam, Flink) and batch pipelines (Airflow, Spark).
- Build and evolve the feature platform, computing features for both streaming and batch, served to rules engines and models under a sub-second budget.
- Establish feature correctness through streaming-versus-batch reconciliation, recomputation tests, and drift monitoring.
- Productionize fraud and identity ML models on Vertex AI and Kubeflow, including automated retraining and champion/challenger promotion.
- Engineer complex risk signals for KYC, AML, and identity verification, transforming multi-vendor data into model-ready features.
- Integrate third-party enrichment providers and manage consortium network failover, caching, and cost.
- Manage the warehouse and modeling layer in BigQuery, including partitioning, staging, and migration strategies.
- Design entity resolution and graph data structures to link customer identifiers across clients.
- Enforce data security and residency requirements, including PII handling and field-level encryption.
PythonSQLGCP+2 more