Design and architect a unified MLOps and forecasting platform on Databricks, leveraging MLflow for model orchestration and Databricks SQL for high-scale data processing.
Build and maintain automated pipelines to handle critical business use cases, specifically:
Developing models that simulate past campaign performance using historical transaction data to forecast future outcomes.
Real-time monitoring and prediction of budget utilization to prevent over/under delivery.
Implementing robust Incrementality Testing framework to quantify the causal impact of advertising exposure.
Architecture and design: Design and evolve the Cardlytics next-generation architecture; establish reference architectures and standards for batch, streaming, ML, and AI/agentic workloads.
AI-ready data platform: Build out the semantic layer, feature store, and data ontology that make Cardlytics data consumable by AI/agentic applications; enable governed, self-service access for both human and machine (LLM/agent) consumers.
