Pomona: Continuous Code Quality Improvement via Small, Automated Changes at Bloomberg

2026-06-04

AI Summary

This work proposes a human-AI collaborative continuous improvement paradigm grounded in the Kaizen philosophy to sustainably enhance code quality and reduce technical debt with minimal risk while preserving engineer trust and productivity. The approach is implemented through Pomona, a lightweight agent-based tool that first identifies and prioritizes code issues—such as lint violations, technical debt, and test gaps—and then generates small pull requests averaging around ten lines of code. In a one-month deployment study, 15 out of 17 automatically generated pull requests were merged, with a median resolution time of under two hours. Furthermore, eight out of ten senior engineers expressed willingness to adopt the tool, demonstrating its practical efficacy and high user acceptance.

Abstract

In this short experience paper, we present Pomona, a lightweight agentic tool that utilises agent skills for continuous automated code quality improvement. Inspired by the philosophy of Kaizen(TM), Pomona automates a cycle of discovery and incremental repair: a Scanning skill identifies improvement tasks (e.g., linting violations, technical debt markers, and test gaps) and prioritises them in a structured backlog, while a Repair skill generates tiny pull requests (PRs) targeting ~10 lines of diff. This human-in-the-loop design enables frequent, low-risk improvements while maintaining engineer trust and productivity and reducing technical debt. We evaluated Pomona through a one-month deployment in a team and a questionnaire distributed to 10 senior engineers. Our preliminary results are promising: 15 of 17 generated PRs were successfully merged with a median time-to-close of under 2 hours. Furthermore, 8/10 of surveyed engineers expressed a desire to adopt Pomona, praising small diff sizes and Pomona's focus on improving code quality. We conclude by discussing actionable insights for researchers and practitioners on strategies for effective agentic deployment in industry.

Problem

• code quality
• technical debt
• automated refactoring
• continuous improvement
• software maintenance

Innovation

• agentic tool
• continuous code improvement
• small pull requests
• technical debt reduction
• human-in-the-loop