Senior Data Engineer

We're looking for a Senior Data Engineer with strong data architecture and system design skills to design, build, and own end-to-end production pipelines powering analytics and decision-making across the org. The role centers on deep Python + SQL expertise, robust transaction-level fact tables, and rigorous data quality and auditability, ideally within a finance, risk, or compliance data domain. Internal candidates familiar with Uber's data platform stack (Piper/uWorc, Databook, uSecret, DSW, SourceGraph, Query Builder, OneETL) will be prioritized.

• Own data architecture and system design decisions for pipelines and data models - grain, partitioning, schema evolution, and scalability tradeoffs. •Take end-to-end ownership of pipelines in production: design, build, deploy, monitor, and operate. •Design and build transaction tables (append-only, immutable event/transaction-grain fact tables) alongside standard fact/dimension models using medallion (bronze/silver/gold) architecture. •Build and support pipelines for finance, risk, or compliance use cases where accuracy, auditability, and data lineage are critical. •Implement data quality and auditability controls: validation checks, reconciliation logic, anomaly detection, and audit trails for every pipeline you own. •Write complex, cross-dialect SQL (MySQL + PostgreSQL): window functions, multi-layered CTEs, CASE-driven logic, COALESCE/NULLIF, type casting, and date/timestamp handling. •Build idempotent reload patterns (DELETE+INSERT), UNION ALL/set operations, and templated (Jinja-style) SQL including handling for late-arriving/corrected transaction records without double-counting. •Orchestrate Airflow-style DAGs (Pipeline/BaseTask, ExternalTaskSensor for cross-pipeline deps) with secure credential handling via uSecret. •Build Python ETL tooling: pandas (CSV ↔ DB), SQLAlchemy + raw drivers (MySQLdb, psycopg2), type hints/dataclasses, class-based pipeline design. •Integrate with Google Drive/Sheets APIs; build YAML-driven pipeline configs; handle Piper staging/user_staging quirks. •Own CI/CD for pipelines via standard Git workflows. •Use Uber's internal stack day-to-day: OneETL for pipeline authoring, Piper/uWorc for scheduling, Databook for lineage, uSecret for credentials, DSW notebooks for CSV/table work, SourceGraph for code search, Query Builder for ad hoc SQL.
• 6+ years in data engineering. •Strong data architecture and system design skills to make sound tradeoffs on modeling, partitioning, and scalability, not just write queries. •Demonstrated end-to-end production ownership of data pipelines from design through deployment, monitoring, and operational support. •Expert-level Python + SQL as the primary toolset. Hive/Hadoop is a nice-to-have only. •Robust data quality and auditability practices - validation, reconciliation, anomaly detection, audit trails. •Experience working with finance, risk, or compliance data - comfort with the heightened accuracy, auditability, and data-lineage requirements typical of these domains. •Proven experience designing and building transaction/fact tables. •Expert SQL (window functions, CTEs, CASE logic, cross-dialect MySQL/PostgreSQL). •Strong Python (pandas, SQLAlchemy, type hints/dataclasses, class-based pipelines). •Airflow-like orchestration experience (Piper/uWorc a plus). •Solid data modeling fundamentals (fact/dim, normalization, medallion architecture). •YAML-based config experience; standard CI/CD and Git workflows.