# AI Coding Tip 032 - Build a Dark Factory Pipeline

DevFeed: [AI Coding Tip 032 - Build a Dark Factory Pipeline](<https://devfeed.tech/articles/ai-coding-tip-032-build-a-dark-factory-pipeline-18222.md>)

Original publisher: [Read original article](<https://maximilianocontieri.com/ai-coding-tip-032-build-a-dark-factory-pipeline>)

Author: Maxi Contieri

Published: 2026-08-16T14:54:12Z

Content type: opinion

Language: en

Sources: [Maximiliano Contieri - Software Design](<https://devfeed.tech/sources/maximiliano-contieri-software-design.md>)

Topics: [ai-coding](<https://devfeed.tech/topics/ai-coding.md>), [pull-requests](<https://devfeed.tech/topics/pull-requests.md>), [coding](<https://devfeed.tech/topics/coding.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding](<https://devfeed.tech/tags/ai-coding.md>), [audit](<https://devfeed.tech/tags/audit.md>), [production](<https://devfeed.tech/tags/production.md>), [pull-request](<https://devfeed.tech/tags/pull-request.md>), [quality-control](<https://devfeed.tech/tags/quality-control.md>), [review](<https://devfeed.tech/tags/review.md>)

## AI overview

This article recommends an automated pull-request pipeline in which one model builds from a specification, a separate model from another vendor or family adversarially verifies the result, and flagged changes are fixed before human review. A defined sample of merged pull requests is randomly audited by humans, with overrides logged as correction signals and merges blocked when either gate is skipped.

## Source excerpt

TL;DR: Run your pipeline like a dark factory: automated, sampled, and policed by an adversarial model. Common Mistake ❌ You let one model write a pull request, then hand the same model (or a suspicio