Professional Experience
Legacy Migration Platform
I built the Python migration pipeline behind Studio Kids Management’s move from legacy Azure SQL records to PostgreSQL. Checkpoint recovery, audit logging, and repeatable writes make long-running transfers recoverable, while data normalization and encryption prepare sensitive agency records for the new platform.
The problem
Replacing the agency’s software also meant carrying forward its existing records. Inconsistent source data, sensitive financial fields, and long-running transfers required a migration process that could be inspected, tested, and resumed after interruption.
How I approached it
I built a batch ETL pipeline with dry-run modes, target-schema validation before writes, checkpoint recovery, and idempotent upserts. Normalization handles inconsistent source records, while encryption and key-management configuration protect sensitive fields. A companion media pipeline converts legacy binary images into web-safe files in Azure Blob Storage.
Inside the build
- Implemented checkpoint-resume flow and dry-run modes for safer long-running migrations.
- Added target schema preflight validation to fail fast before writes.
- Built normalization logic for inconsistent source data and idempotent upsert patterns.
- Secured PII fields with encryption and key-management aware configuration.