Human-In-the-Loop Software Development Agents

2024-11-19

arXiv.org

AI Summary

Current LLM-based programming approaches lack real-time human-AI collaboration and validation in authentic environments, resulting in insufficient code quality assurance. This paper introduces HULA—a novel, industrial-grade, closed-loop human-AI collaborative LLM agent system. HULA deeply integrates engineers’ real-time interventions throughout coding planning, generation, and verification via a Jira plugin, leveraging a multi-agent architecture, interactive prompt engineering, and progressive task validation. In internal deployment, HULA significantly reduces time-to-first-commit and implementation latency for simple tasks; engineers report markedly decreased development time and cognitive load. Although code quality in complex scenarios remains an area for improvement, this work pioneers the systematic integration of structured human feedback across the entire software development lifecycle—addressing a critical gap in deploying LLMs within real-world engineering environments.

Technology Category

• Humans and AI: Human-in-the-loop Machine Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Hardware-aware ML

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Recently, Large Language Models (LLMs)-based multi-agent paradigms for software engineering are introduced to automatically resolve software development tasks (e.g., from a given issue to source code). However, existing work is evaluated based on historical benchmark datasets, rarely considers human feedback at each stage of the automated software development process, and has not been deployed in practice. In this paper, we introduce a Human-in-the-loop LLM-based Agents framework (HULA) for software development that allows software engineers to refine and guide LLMs when generating coding plans and source code for a given task. We design, implement, and deploy the HULA framework into Atlassian JIRA for internal uses. Through a multi-stage evaluation of the HULA framework, Atlassian software engineers perceive that HULA can minimize the overall development time and effort, especially in initiating a coding plan and writing code for straightforward tasks. On the other hand, challenges around code quality remain a concern in some cases. We draw lessons learned and discuss opportunities for future work, which will pave the way for the advancement of LLM-based agents in software development.

Problem

• Human-in-the-loop
• Code Quality
• Programming Tasks

Innovation

• HULA framework
• real-time code generation
• user-adjustable advanced language models