Engineering Trustworthy LLM Systems

We explore how to engineer reliable and effective Large Language Model (LLM) systems for real-world applications. With the rise of generative AI and foundation models, organizations face new challenges in integrating LLMs into their software engineering workflows.

We investigate quality characteristics of LLM systems and develop solutions to ensure trustworthiness, including a reference architecture for Retrieval-Augmented Generation (RAG) applications. Key focus areas include prompt engineering, validation frameworks, and LLMOps—covering model selection, document retrieval, interface design, and system evolution.
Fontys collaborates with industry partners to apply LLMs in software engineering (LLM4SE) and improve engineering processes using AI (SE4LLM). The goal is to operationalize AI-augmented software engineering in existing environments and educate future engineers for the GenAI era.

Companies are invited to co-develop LLM systems, contribute to the growing body of knowledge, and shape the future of trustworthy AI in software development.