The SILAS workshop brings together researchers, practitioners, and policymakers to explore how sustainability intersects with software engineering practices and large language models (LLMs). We encourage participation regardless of whether you publish at SILAS, as we appreciate the value of diverse insights and active dialogue.

LLMs are rapidly becoming part of everyday software engineering workflows. Their integration into tasks such as requirements, design, implementation, testing, and maintenance—combined with large CI/CD pipelines—raises important questions about energy use and resource consumption. Understanding and reducing this footprint is increasingly urgent.

The second edition of SILAS will be held fully on-site and features a keynote (TBA), paper presentations, and structured roundtable discussions aimed at identifying challenges and opportunities for more sustainable LLM-assisted software engineering.

Topics of interest include, but are not limited to:

• empirical studies on the environmental or resource impact of LLM-assisted software engineering
• datasets, benchmarks, or measurement methodologies for evaluating LLM efficiency
• tools, techniques, and practices for reducing the resource footprint of LLM-assisted tasks across the software engineering lifecycle
• profiling, analysis, or characterization of LLM behavior in SE contexts
• efficient and repeatable prompting strategies, including resource-efficient code generation
• industrial experiences, case studies, or demonstrations

Submission Guidelines

Accepted papers will be published in the ICT4S 2026 Workshops Proceedings (CEUR)

Submissions must use the 1-column CEUR template (ODT or Overleaf).

We invite three categories of papers:

• Research papers (10-15 standard pages) present mature and well-motivated work that advances the understanding of sustainable LLM-assisted software engineering. These submissions will be evaluated for relevance, novelty, soundness, and readability
• New Ideas and Emerging Results (NIER) papers (5–9 standard pages) present early-stage, promising, or provocative ideas that have the potential to open new research directions in sustainable LLM-assisted software engineering. These submissions may describe novel concepts, under-explored challenges, initial observations, or emerging results that are not yet mature enough for a full research paper. The emphasis is on insight, originality, and the capacity to stimulate discussion. NIER papers will be evaluated for their relevance, novelty, potential impact, and clarity.
• Industrial application & demonstration papers (5–9 standard pages) present practical experiences, applications, tools, or demonstrations related to sustainable LLM-assisted software engineering. Submissions may describe real-world deployments, lessons learned, performance observations, or applied techniques used in industry or public-sector contexts. These papers should clearly convey the practical relevance and insights gained from implementation or use. Industrial and Demonstration papers will be evaluated for their relevance, practical value, and clarity of presentation.

All submissions must be original and peer-reviewed by a program committee of international experts.

Organizing committee

• Pepijn de Reus, University of Amsterdam
• Maja H. Kirkeby, Roskilde University
• Fernando Castor, University of Twente
• Ana Oprescu, Universiteit van Amsterdam.
• João Paulo Fernandes, New York University Abu Dhabi.

Venue

The conference will be held in Bern, Switzerland together with ICT4S 2026.

Contact

All questions about submissions should be emailed to p.dereus@uva.nl