Senior Principal, Data Product Engineering

See how Epsilon powers connected marketing by unlocking the full potential of data and identity.

Epsilon's transactional data products turn purchase and card-transaction signals into governed datasets that our audience, measurement, and client reporting teams build on. We're hiring a Senior Principal Engineer to own the quality and reliability of those datasets on our lakehouse platform.

This is a hands-on individual contributor role. You'll be in the pipelines and in the SQL, accountable for data that goes directly into client work - where correctness and timeliness are product requirements. It suits someone who is happiest close to the data, unbothered by a messy real-world domain, and treats doing the same fix twice as a bug.

• Empower with accountability:
  You own your data products end to end - ingestion through publication - and the engineering standards the rest of the team builds against.
• Innovate with purpose:
  You'd rather eliminate a class of problem than resolve an instance of it.
• Work together to win together:
  This role reports to the Senior Director, Data Product Engineering.

• Quality you can prove. Build the automated QC layer - completeness, distribution, referential integrity, anomaly detection - and turn the signal into a small number of alerts people actually trust.
• Reliability you own. Diagnose failures and discrepancies across large, high-cardinality data, and set freshness and correctness expectations you can demonstrate you're meeting.
• Toil that disappears. Replace recurring manual operations with automation and self-service tooling, and make compute proportional to what actually changed rather than to the size of the dataset.
• A platform that keeps improving. Evolve the Spark and SQL pipelines you own - tune refresh strategy and cost, retire the scaffolding that has outlived its purpose, and make each new source cheaper to onboard than the last.
• A higher bar. Set the standards other engineers follow - shared libraries, automated testing, code review, declarative deployment - and mentor the engineers and analysts around you.

• 10+ years hands-on with production data, across data engineering, data operations, or data platform work
• Expert SQL, including performance tuning where a bad query plan costs hours and real money rather than seconds, plus strong production Python (PySpark preferred)
• A track record designing data quality, validation, monitoring, or reconciliation systems that measurably reduced defects or manual effort
• Real accountability for production batch pipelines, not only experience building them
• Large-scale distributed platforms (Databricks, Spark, or equivalent) and relational or MPP warehouses (Greenplum, Postgres, Snowflake, Redshift, or similar)
• Software engineering practice applied to data work: version control, code review, automated testing, CI/CD
• Comfort with messy, high-cardinality data - matching and deduplication included - and healthy skepticism toward data that looks too clean
• The ability to explain how an upstream change ripples into a downstream product to people who aren't engineers
• Bachelor's degree in a technical field, or equivalent practical experience

• Transaction, payments, card, retail POS, loyalty, or retail measurement data experience
• Delta Lake or equivalent, governed catalogs, incremental ingestion, and cluster or warehouse cost tuning
• Workflow orchestration at scale (Airflow, Databricks Jobs) and containerized deployment
• PII handling, pseudonymized identifiers, or work in regulated or clean-room environments
• Lightweight internal tooling (Flask or similar) for operational interfaces