QA Engineer - Data (Contract)
QA Analyst - Microsoft Fabric / Data Lakehouse Contract • Independent Contributor*Must be in Los Angeles area (hybrid)*No 3rd parties - No Sponsorship available - No C2CWe're looking for a seasoned data QA specialist to serve as the sole, independent test function for a Microsoft Fabric medallion architecture build. You'll own the entire QA lifecycle - test case design through gate sign-off - validating data across Bronze, Silver, and Gold layers with no day-to-day direction.
What you'll do:
- Author and maintain the full test-case library and unit-test catalog across all medallion layers, mapped to source-to-target mappings and grain statements
- Write repeatable SQL checks for column/type contracts, key uniqueness, grain, layer-to-layer reconciliation, referential integrity, and pipeline idempotence
- Produce test plans and evidence packs at each delivery gate (freeze, UAT, production promotion, hypercare) with written go/no-go recommendations and residual risk summaries
- Validate semantic model bindings, relationships, measures, and row-level security; confirm reports consume certified models rather than raw lakehouses
- Log defects with layer attribution and severity - complete enough for engineers to act without a follow-up meeting
- Confirm CI and data-quality notebooks pass before deployment; verify promotion order, rollback readiness, and flag any exposed credentials as security defects
- Deliver a weekly written status covering coverage, pass rates, open defects, and risks to the next gate; escalate blockers same-day
What we're looking for:
- Hands-on data, ETL/ELT, or BI QA experience reconciling data across layers - not UI-only testing
- Demonstrated ability to author test cases and test plans from mappings, data dictionaries, and grain statements
- Strong SQL - joins, aggregation, window functions, set comparison, duplicate and null profiling, and layer-to-layer reconciliation queries written unaided
- Hands-on Microsoft Fabric and/or Power BI experience across lakehouses, notebooks, semantic models, and published reports in dev/test/prod workspaces
- Solid understanding of dimensional modeling - grain, surrogate vs. business keys, facts vs. dimensions, SCD basics
- Proven independent delivery in a small or single-QA environment with strong written and verbal communication
- Strict discipline with sensitive data - masked samples, no PII/PHI in test artifacts, no secrets in documents
Nice to have:
- Fabric medallion delivery and Direct Lake semantic model experience
- Health plan, claims, pension, or retirement benefits data background
- PySpark or notebook-based test assertions, or experience with frameworks like Great Expectations
- Azure DevOps or Git with deployment pipelines
- Legacy on-premises SQL Server source experience