2026· International Conference on Software and Data Technologies· pp. 246-253· 0 citations· 24 references
Computer Science
TL;DR
The case suggests that generative AI is especially useful when requirements are only partially formalized, yet objective feedback from tests, benchmarks, and model quality metrics is available, and the results suggest that AI-augmented development is a relevant topic for scientific software engineering.
Abstract
: Generative AI has intensified interest in Automated Software Engineering, but most current evidence still concerns short, local coding tasks rather than the broader software engineering lifecycle. This paper presents a structured experience report on the development of a Python package for interpretable multiclass classification that hosts heterogeneous estimator families behind a shared interface. We analyze how AI assistance contributed to four recurring engineering activities—cross-language transfer, paper-to-code implementation, benchmark-driven algorithm refinement, and packaging/refactoring work—and how a small set of machine-checkable contracts acted as a cross-cutting enabler that made AI-assisted changes safer to accept. The contribution of the paper is threefold: a concrete research-software case from interpretable machine learning, a task-oriented account of where AI created practical leverage, and a set of lessons on planning, architecture, and validation for trustworthy use. The case suggests that generative AI is especially useful when requirements are only partially formalized, yet objective feedback from tests, benchmarks, and model quality metrics is available. At the same time, the study indicates that human oversight remains essential for semantic correctness, experimental validity, and maintainable software architecture. Overall, the results suggest that AI-augmented development is a relevant topic for scientific software engineering.
Enterprise software teams report a frustrating inconsistency: the same AI coding assistants producing compelling results in controlled studies return uneven outcomes in production. The problem is architectural, not technological. A capable AI coding assistant deployed without adequate codebase context, integrated at only one SDLC touchpoint, and missing any mechanism through which developers calibrate their trust in its suggestions will underperform relative to its documented potential regardless of its underlying model quality. This article argues that three integration conditions determine whether enterprise deployments match the productivity outcomes controlled experiments document: context fidelity, the degree to which the assistant's knowledge reflects the actual enterprise codebase, workflow embeddedness, the breadth of development lifecycle touchpoints at which assistance is available and trust calibration architecture, the instrumentation through which developers form accurate expectations of suggestion quality. Design patterns for each condition are specified retrieval-augmented generation for context management, multi-surface integration strategies for embeddedness, and feedback loop instrumentation for trust calibration, alongside a risk classification framework addressing hallucination, security vulnerability injection, and over-reliance degradation. A governance model closing the continuous improvement cycle between developer feedback, quality measurement, and policy adjustment completes the specification. The analysis draws on the author's experience architecting enterprise AI automation systems across global software operations, and on a body of empirical literature whose collective finding is that tool capability and integration architecture are equally necessary conditions for sustainable productivity gains.
Rajendar Reddy Sama· East African Journal of Info...· 0 citations
AI-assisted development tools enable software engineers to generate implementations at substantially higher speed and volume than in traditional workflows, yet relatively little is known about how existing guardrails evolve in response.
This review traces how quality engineering has changed, from rule-based automation to self-adjusting test frameworks, and looks at the technology behind Autonomous Quality Agents: Large Language Models that generate code from requirements, Computer Vision that handles visual regression, and Reinforcement Learning that drives exploratory testing.
Vanshita Agarwal· International journal for ad...· 0 citations
Generative AI tools such as GitHub Copilot, ChatGPT, and coding agents have rapidly become part of everyday software development, yet little is known about how mainstream open source communities discuss them in practice. This paper presents a longitudinal analysis of generative-AI-related discussions in the Visual Studio Code (VS Code) GitHub repository, using 43,806 candidate issues created between January 2021 and June 2026. To improve corpus quality, we combined keyword retrieval with semantic relevance filtering, yielding a filtered corpus of 25,227 AI-related issues. We applied BERTopic to the retrieved corpus to identify discussion topics, using the filtered corpus for theme validation and a robustness re-clustering, and analyzed their evolution over time using monthly prevalence and Mann-Kendall trend tests. The results show that developer discussions are dominated by practical concerns regarding the operation of AI-assisted development environments, including agent management, configuration, reliability, authentication, and billing, whereas risks frequently emphasized in survey-based studies, such as hallucination and licensing, rarely surface in this venue. This suggests that discussions of generative AI in the VS Code issue tracker primarily focus on operational aspects of AI-assisted software development. Furthermore, discussions evolved from AI-assisted code completion toward conversational and agent-based development, reflecting the increasing integration of generative AI into software development workflows. These findings suggest that GitHub Issues provide a practical, workflow-oriented perspective on generative AI that complements survey-based studies of developer perceptions.
Panida Rumriankit, Akito Monden, Hiroki Inayoshi et al.· 0 citations
Thanks to advancements in generative AI and Large Language Models (LLMs), the last five years have seen exponential growth in the adoption of AI-Assisted software engineering across many industries. Simple developer tools used for code completion, static analysis and syntax linting have been augmented by semi-autonomous agents, able to contribute to a broad set of Software Engineering practices including code generation, documentation generation, test creation, refactoring, architectural assistance, enhanced Integrated Developer Environment (IDE) tools and analysis and verification. The way we develop, analyze and test software is rapidly changing. Recent industry reports show that at least 84% of professional software developers already use AI coding assistants regularly, with 52% of developers reporting that AI tools have had a positive effect on their productivity. At GMTO, we stand to benefit from these technologies and adopting industry best practices. However, it needs to be done in a way that carefully considers our unique concerns and risks. These include code quality, maintainability and technical debt; scientific and engineering integrity, governance, trust and over-reliance; and ethical, talent and workforce issues. In this paper we describe how we are starting to utilize AI-assisted software engineering at GMTO, why these common industry concerns matter for Observatory software teams, and future plans.
M. Cox, J. Filgueira, Chien Peng et al.· Astronomical Telescopes + In...· 0 citations
A three-level taxonomy inspired by autonomous driving that distinguishes degrees of autonomy along a roadmap from today’s AI-assisted development workflows to fully autonomous software development in which AI systems autonomously identify demands and design, implement, verify, and maintain software without human oversight is introduced.
Hao Wang, Ruijie Meng, Zhe Ye et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
