arXiv subjects

Petrus Lipsanen

Publications and source records attributed to Petrus Lipsanen.

3 recordsLinked to original sources

arXiv2026

TimeArch: A Multi-Agent Approach for Software Architecture Design

Generative AI is making software implementation faster, but not necessarily better architected. Most current GenAI coding workflows move directly from prompts to code, while the architectural reasoning that should connect requirements, quality attributes, stakeholders, design decisions, and implementation remains weakly supported. Architectural styles, system decomposition, tradeoff analysis, quality-attribute decisions, and stakeholder concerns are often left implicit or handled outside the AI-supported workflow. This creates a missing architectural layer between requirements and code, limiting traceability, validation, and governance in GenAI-enabled software development. This paper presents TimeArch, a multi-agent approach that makes this architectural layer explicit. TimeArch structures software architecture design as a governed lifecycle rather than a one-shot generation task. It combines agent-based reasoning with retrieval-augmented grounding to support requirement definition, architecture design, validation and assurance, and delivery and evolution. The approach supports requirement capture, architectural driver identification, design reasoning, validation, and decision documentation within a traceable process. We demonstrate TimeArch through a working prototype that operationalizes the proposed approach and assess its feasibility using representative software architecture design cases from multiple domains with heterogeneous inputs. As future work, we plan to conduct comprehensive industrial case studies to evaluate the adoption and use of TimeArch in both new and existing software projects. A TimeArch demo is available at: https://sda-assistant.com/. Keywords: AI-native software architecture; requirements-driven software architecture; architectural reasoning; architecture traceability; architecture validation

cs.SE↗

arXiv2026

Shift-Up: A Framework for Software Engineering Guardrails in AI-native Software Development -- Initial Findings

Generative AI (GenAI) is reshaping software engineering by shifting development from manual coding toward agent-driven implementation. While vibe coding promises rapid prototyping, it often suffers from architectural drift, limited traceability, and reduced maintainability. Applying the design science research (DSR) methodology, this paper proposes Shift-Up, a framework that reinterprets established software engineering practices, like executable requirements (BDD), architectural modeling (C4), and architecture decision records (ADRs), as structural guardrails for GenAI-native development. Preliminary findings from our exploratory evaluation compare unstructured vibe coding, structured prompt engineering, and the Shift-Up approach in the development of a web application. These findings indicate that embedding machine-readable requirements and architectural artifacts stabilizes agent behavior, reduces implementation drift, and shifts human effort toward higher-level design and validation activities. The results suggest that traditional software engineering artifacts can serve as effective control mechanisms in AI-assisted development.

cs.SE↗

arXiv2025

Towards Shift-Up: A Framework and a Prestudy on High-Value Activities in GenAI Native Software Development

Generative AI (GenAI) has significantly influenced software engineering. Associated tools have created a shift in software engineering, where specialized agents, based on user-provided prompts, are replacing human developers. In this paper, we propose a framework for GenAI native development that we call \textit{shift-up}, which helps software teams focus on high-value work while being supported by GenAI. Furthermore, we also present a preliminary study testing these ideas with current GenAI tools. Towards the end of the paper, we propose future research goals to study shift-up in more detail.

cs.SE↗