Data Migration & Batch Engineer
Role details
Job location
Tech stack
Job description
· Own schema-drift detection and reconciliation (Red Gate / SQL comparison) - the deterministic ground truth on which every downstream decision rests.
· Build the deterministic rule ladder for table and job dispositions: staging / truncate-and-reload detection, dead-table detection, target-equivalent matching, straight-through repoint, duplicate-job merge.
· Own the ID-collision engine: a full (never sampled) overlap scan across every shared ID space; classification into true collision, phantom and historical overlap; analysis of the resolution strategy; and a non-bypassable validation harness on every script that touches an ID column.
· Own data reconciliation: row counts, column checksums, referential integrity, stored-procedure output parity, business-key parity; staging promotion gates and validated rollback scripts.
· Batch: build the dependency graph, blast-radius (butterfly) scoring, Simple/Medium/Complex classification, and the migration approach per job.
· Own the SLA validation harness - proving every migrated job and every SLA chain runs within its window under combined nightly volume.
· Emit the downstream artifacts: migration scripts, job modifications, rollback scripts, reconciliation specifications.
Requirements
Experience: 8+ years in relational data migration and/or ETL engineering, with cutovers delivered to production. You own the deterministic core of the database and batch workstreams. Two systems have minted identifiers independently for many years across about a million accounts; merging them incorrectly mis-maps an investor''s assets. This seat is where correctness is defended., · SQL Server (2024) - deep: DDL, stored procedures, indexing, referential integrity, query performance.
· Relational data migration at scale - Red Gate, AWS DMS or equivalent; cutover, reconciliation and rollback delivered in production.
· Identity and key-collision resolution - surrogate versus natural keys, reconciliation and historical-mapping tables, preserving referential integrity under a merge. Rare, and highly valued for this role.
· Informatica PowerCenter; ETL dependency analysis; SLA and batch-window modelling.
· Python and strong SQL; data-quality and validation frameworks.
· The discipline to keep correctness-critical steps deterministic while working alongside an LLM reasoning layer.
Working knowledge
· Informatica IDMC and/or Apache Airflow; PySpark.
· Graph queries; handling of regulated / PII data.