Senior Data Engineer
Job Description
NSPR Law Services LLC is seeking a Senior Data Engineer to support production data reliability across transactional PostgreSQL and an Iceberg/Athena lakehouse. This onsite role in San Juan, PR focuses on building ETL and dbt transformation layers, strengthening identity-resolution and data-quality controls, and leading a safe migration away from an unreliable replication path. The work is centered on traceability and reconciliation, so downstream systems can trust the data.
Salary: USD 100,000 - 110,000 per year
Experience: 5+ years
Work location: In person
Benefits
- Dental insurance
- Health insurance
- Vision insurance
- Life insurance
- Paid time off
- Flexible schedule
- Employee assistance program
- Professional development assistance
- Referral program
- Retirement plan
What You’ll Do
- Design and build production ETL pipelines and the dbt transformation layer with tested transformations for operational PostgreSQL and analytical models in the Iceberg/Athena lakehouse.
- Reverse-engineer undocumented business logic from legacy schemas, encoding and testing transformations instead of copying tables.
- Own the target PostgreSQL schema for the shared data layer used by every Bronson product.
- Build an identity-resolution pipeline using deterministic and probabilistic matching, confidence scoring, a manual-resolution queue, and measured false-merge and false-split rates.
- Implement data-quality and lineage controls across both layers, including schema-change detection, freshness monitoring, and traceability from source records to product-facing models.
- Own parallel run reconciliation, keeping legacy systems and the new layer aligned with documented discrepancies until the old side is switched off.
What You’ll Need
- 5+ years of professional data engineering experience.
- Built and owned a dbt (or equivalent) transformation layer in production, including tests and CI covering transformations, schema changes, and pipeline code.
- Led a migration from legacy on-prem databases to a modern platform using parallel runs, with end-to-end source-to-target reconciliation and documented, approved exceptions.
- Deep SQL and PostgreSQL experience, including fluent work across other teams’ schemas.
- Change-data-capture experience (Debezium, DMS, or similar).
- Strong attention to detail and patience while verifying legacy logic that exposes issues over time.
- Comfort communicating with product and business stakeholders to validate undocumented rules, resolve discrepancies, and document source-of-truth decisions.
Technical Environment
- Transactional core on managed PostgreSQL with schema-per-domain and a transactional outbox.
- Analytics lakehouse on object storage using Iceberg tables, a SQL query engine, and dbt transformations.
- Infrastructure as code (TypeScript) and GitHub Actions CI/CD.
- Python for data tooling; observability with OpenTelemetry where applicable.
- Legacy sources include on-prem SQL Server and MySQL.
- Environments: dev (synthetic data), staging (de-identified), and production (the only PHI environment).
Key Technical Challenges
- The existing replication path is no longer reliable, requiring you to choose a replacement change-data-capture approach (DMS-class vs Debezium-class), defend it, and own it in production.
- Identity matching can become a legal issue, so matching must include measured false-merge and false-split rates and a manual-resolution queue.
- Each pipeline reconciles against its legacy source with documented discrepancies, with reconciliation treated as part of the pipeline.
- Backfills and replays must be safe: re-running the same load must never produce different results.
Technologies: PostgreSQL, dbt, Iceberg, Athena, TypeScript, GitHub Actions CI/CD, Python, OpenTelemetry, SQL Server, MySQL