Agentic Software Engineering
We investigate autonomous and semi-autonomous AI agents throughout the software lifecycle, with explicit responsibilities, useful context and verifiable outcomes.
We research new approaches, methods, tools and technologies for building high-quality, secure and trustworthy information solutions. Our focus connects AI4SE, agentic software engineering, software quality, DevOps/DevSecOps, modern architectures and user experience.
We investigate autonomous and semi-autonomous AI agents throughout the software lifecycle, with explicit responsibilities, useful context and verifiable outcomes.
We study how artificial intelligence and generative AI support software engineering. Systematic use and evaluation matter more than adopting the latest tool.
We connect software and process metrics with empirical evaluation. Our goal is trustworthy, maintainable and sustainable software, supported by evidence.
We investigate automation throughout the software lifecycle, including the integration of AI agents into secure, observable development and delivery pipelines.
We research architectures for adaptable information systems, connecting architectural decisions with reusable solutions and the realities of distributed environments.
We examine how people experience, use and accept technology. Accessibility, inclusion and human–agent interaction are part of responsible software engineering.
Systematic. Measurable. Reusable. Research becomes useful when we can explain how, where and why an approach works.
Ask a useful question.
Test under defined conditions.
Find what can be repeated.
Learn in real environments.
Bring the knowledge to work.
We do not study AI agents in isolation. We examine how agentic approaches interact with development processes, architecture, software quality, DevOps/DevSecOps, engineering governance, teams, users and organizations. Our focus is systematic, measurable and reusable engineering approaches.
We design and evaluate workflows involving coding agents and specialized AI agents, including responsibilities, context and verification.
We systematically evaluate coding agents, foundation models, agent frameworks, development environments, orchestration approaches and engineering workflows.
We identify, document and evaluate reusable interaction, orchestration and development patterns, engineering practices and verification approaches.
We investigate AGENTS.md, project context, specifications, plans, architecture documentation, coding standards, skills, acceptance criteria and verification scripts.
Controlled and applied experiments examine correctness, quality, maintainability, effort, productivity, reliability, autonomy and context requirements.
We develop prototypes and industry pilots to investigate engineering questions in realistic environments.
We connect findings with practice, education, workshops, prototypes, pilots and good practices. Transfer also brings new research questions back to the lab.