His software factory runs on two loops and a rocket emoji

2026-07-15 · Source: How I AI · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Fundamental Awareness, quick

Summary

A software factory system, designed for the SaaS Henry Intelligent Machines, operates through a two-loop process: a build loop and a review loop. The process begins with a "morning build" where the user interacts with a cloud system, answering questions that generate a list of development tasks. Subsequently, a dedicated build loop agent iteratively executes these tasks, constructing the necessary software components. Following this, a separate review loop agent meticulously examines the built tasks, identifying and rectifying any code issues or discrepancies. Upon successful review, the system notifies the user via Slack, presenting the completed and validated work. The user then triggers a final merge into the "Henry loop" by simply leaving a rocket emoji in Slack, completing the development cycle.

Key takeaway

For MLOps Engineers or Software Engineers seeking to streamline their development workflow, consider implementing a similar two-loop automation system. By separating iterative build tasks from an independent review agent, you can significantly accelerate development cycles and ensure code quality. Integrate a simple, low-friction approval mechanism, like a Slack emoji, to quickly merge validated work, reducing manual overhead and improving deployment velocity.

Key insights

The system automates software development and review via distinct agent-driven loops, triggered by user input and a simple emoji.

Principles

• Automate iterative task execution.
• Separate build from review agents.
• Use simple triggers for merges.

Method

The process involves a "morning build" for task generation, an agent-driven "build loop" for execution, a "review loop" for quality assurance, and a Slack-based "rocket emoji" trigger for merging.

In practice

• Implement agent-based build automation.
• Establish a separate code review agent.
• Integrate Slack for merge approvals.

Topics

• Software Automation
• Development Workflow
• Agent-based Systems
• CI/CD
• Code Review Automation
• Slack Integration

Best for: AI Engineer, Software Engineer, MLOps Engineer

Related on AIssential

• So we're doing loops now — Agent loops enable autonomous, continuous goal-seeking workflows, transforming discrete interactions into persistent operations.
• Bye-Bye Manual: Making Development Simple With a 2-Agent Pipeline — A 2-agent AI pipeline, "The Architect" and "The Executor," automates development via file-based handoffs, simplifying multi-role AI sessions.
• Shipping huggingface_hub every week with AI, open tools, and a human in the loop — Automating software releases with AI and open tools requires deterministic validation and human oversight for trustworthiness.
• AIEWF Daily Dispatch: Loops, Software Factories & Forward Deployed Engineers — AI agents, orchestrated through iterative loops, are foundational to the emerging software factory paradigm for software development.
• Building verification loops in Claude Code with skills — Automating code verification with Claude Code skills enables AI agents to self-correct, closing feedback loops.
• Stop Prompting. Start Writing Loops — META LOOPS offer a token-efficient workflow design, moving beyond simple prompting to optimize AI system performance.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by How I AI.