Split a huge legacy codebase into modules with an agent
- Level Advanced
- Setup1–3 hrs (estimated)
- Evidence Anecdotal
- CostNot stated in source
- No-codeNo
What they did
Nubank used Devin to break up an eight-year-old ETL monolith with millions of lines of code. The work was split into many small sub-tasks, and Devin was fine-tuned on earlier migrations that engineers had done by hand.
How it went
Cognition reports that the fine-tuning doubled task completion and cut each sub-task from about 40 minutes to 10, with an 8–12x efficiency gain, more than 20x cost savings, and the project cut from an estimated 18 months to weeks.
Worth knowing
- Vendor case study.
- It worked because the job was broken into many similar, repeatable sub-tasks, with human-made examples to learn from.
Prompt
Cleaned version rewritten from the source — not a direct quote.
Migrate [module] out of the monolith following the pattern in [example PRs]. Keep behavior identical, run [tests], and open one PR per [unit] with a summary of what moved.
Recommended
Try it in Grok Bot
- Copy the prompt (button below).
- Open Grok Bot chat and paste it. Replace any [bracketed] placeholders.
- Connect the listed tools (as needed) as connectors when Grok Bot asks.
- If the task repeats, ask: "Make this a daily routine."
Alternatives: the same prompt works in Devin, Claude, ChatGPT Agent, OpenClaw.
Set it up
Stack: Devin
- Read the original source for full context.
- Connect the tools your agent needs.
- Paste the prompt and replace [bracketed] placeholders.
- Run once manually and check the output before scheduling it.
Generic suggested steps, not from the source.
Evidence
Vendor case study. It worked because the job was broken into many similar, repeatable sub-tasks, with human-made examples to learn from
Source: Web · Verified by us: not yet · Report broken or outdated