# How did Nubank use Devin to migrate a multi-million-line ETL monolith?

> Nubank delegated the migration of its 8-year-old, multi-million-line core ETL monolith to Devin. Cognition reports a 12x efficiency improvement in engineering hours saved and over 20x cost savings.

- **Evidence grade:** B — Named organization with a quantified outcome, from the vendor or the organization's own announcement.
- **Organization:** Nubank (digital bank)
- **Industry:** Banking & Financial Services — Fintech bank (Latin America)
- **Roles:** Software Engineering
- **Tools:** Devin (Cognition)
- **Status:** production · **Agent pattern:** autonomous-agent · **Autonomy:** 2/4
- **Last verified:** 2026-10-17

## Reported outcomes
- engineering efficiency: 12x improvement in engineering hours saved [1] (self-reported)
- cost savings: over 20x [1] (self-reported)

## How it works
1. Engineers delegate migration sub-tasks to Devin.
2. Engineers review and merge Devin's changes.

## Risks & compliance
- Agent-written code requires human review and testing
- Vendor-published results; not independently audited

## Sources
1. Cognition (Devin), "How Nubank refactors millions of lines of code to improve engineering efficiency with Devin" (undated; accessed 2026-10-17) <https://devin.ai/customers/nubank>
   > "engineers were able to delegate Devin to handle their migrations and achieve a 12x efficiency improvement in terms of engineering hours saved, and over 20x cost savings."

Canonical: https://agentcasefile.com/cases/nubank-devin-etl-migration/ · License: CC BY 4.0 (summary) · Source: Agent Casefile