• Define partner onboarding and clean room architecture patterns across Snowflake, LiveRamp, and Databricks that are secure, scalable, and repeatable • Configure and manage partner-specific clean room environments; deploy and manage Python-based libraries within the platform ecosystem • Establish and maintain MLOps practices, including model serving, monitoring, and pipeline orchestration for AI/ML features deployed within the platform ecosystem • Own design and enforcement of granular RBAC policies and least-privilege service accounts • Serve as the technical lead for onboarding new partners, implementing privacy-preserving controls (e.g., aggregation thresholds and anonymization techniques) • Design, build, and operate scalable ELT pipelines using Snowpark and/or PySpark and advanced SQL to provision Gold datasets • Implement and evolve identity resolution logic mapping internal data to 3P identifiers (including LUIDs, RampIDs, TransUnion IDs), ensuring privacy-safe practices • Design and operate scalable data architectures across Snowflake and Databricks supporting batch and near real-time processing patterns • Build robust automated checks (e.g., Great Expectations or custom SQL assertions) and define quality standards to detect schema drift, null rate spikes, and volume anomalies • Lead performance optimization across platforms (query tuning, caching, incremental processing) and define and implement query tagging and chargeback models for accurate cost attribution • Establish monitoring, alerting, runbooks, and standard operating procedures to improve platform reliability and reduce incident time-to-resolution • Validate that output data adheres to privacy and business requirements, and define test strategies for partner-facing releases • Serve as the escalation point for diagnosing connection failures, data discrepancies, or latency issues with partner technical teams • Design and build internal AI agents (using frameworks like LangChain, Snowflake Cortex) and mentor other engineers through code reviews, design discussions, and operational best practices