Split a huge legacy ETL monolith into smaller jobs
A large bank used an autonomous coding agent to migrate a multi-million-line ETL monolith into sub-modules; the vendor reports about 12x engineering-hour efficiency and over 20x cost savings.
Steps
- Pick one repetitive migration pattern and document it
- Let the agent do a few migrations; engineers review each
- Scale only once review effort per change drops
Starter prompt
Migrate the job in etl/legacy/customer_daily.py to the new module layout described in MIGRATION.md. Keep behaviour identical, add a comparison test on the sample data, and open a PR.
Guardrail: Vendor-reported numbers; measure your own.
Prompts are our own examples. Results from sources are as reported by their authors.