Dr Klaudia Dussa-Zieger

ISTQB® President, Germany

Dr. Klaudia Dussa-Zieger is the team leader responsible for consulting at imbus. She is particularly interested in test management, the continuous improvement of the test process and the professional training and further education of testers.

For more than 20 years, she has been active as a trainer for the ISTQB® Certified Tester Foundation and Advanced Level and also as a lecturer for software testing at the University of Erlangen-Nuremberg.

Since March 2009, Klaudia Dussa-Zieger has been chairwoman of the DIN working committee "System and Software Engineering" and is actively involved in the creation of standards at international level. For more than 10 years she is member of the German Testing Board (GTB). Currently, she is the President of the ISTQB, where she heads the AI Taskforce.

LLMs and test automation

After a short introduction to Generative AI, the lecture will focus on large language models (LLM). We will explain what a large language model is. Basic terms such as tokens and word embeddings are explained and the functionality of an LLM and a reasoning model is presented. The characteristics as well as the advantages and disadvantages of LLMs are clearly illustrated.

The topics of prompt engineering, retrieval augmented generation (RAG) and fine tuning are then examined and the differences between the various approaches are explained.

After this introduction, it is shown how LLMs can be used in the test process and in test automation - from test case specification to support in test automation. A concrete project example illustrates AI-supported test case specification in a regulated environment. The Claude 3.5 and Llama 3.0 language models are used in combination with a RAG approach. The various work steps are described in detail and it is explained how testers can work effectively with the AI application.

Rik Marselis

Sogeti, Nederland

Rik Marselis is principal quality consultant at Sogeti in the Netherlands. He is a highly regarded presenter, trainer, author, consultant and coach who supported many organizations and people in improving their quality engineering & testing practice by providing useful tools & checklists, practical support and having in-depth discussions. His presentations are always appreciated for their liveliness, his ability to keep the talks serious but light, and his use of practical examples with humorous comparisons.

Rik is an accredited trainer for TMAP, ISTQB and TPI certification training courses, but also he has created and delivered many bespoke workshops and training courses. For example, on quality engineering for Intelligent Machines and DevOps. He is the chairman of the TMAP special interest group.

Rik is a fellow of Sogeti’s R&D network SogetiLabs. These R&D activities result in presentations, books, white-papers, articles, podcasts and blogs about IT in general and quality engineering & testing in particular. He is a co-author of the TMAP book “Quality for DevOps teams” and contributor to the www.TMAP.net body of knowledge for quality engineering & testing.

In 2022 Rik received the ISTQB Software Testing Excellence Award and the EuroSTAR Best Tutorial Award.

Five reasons why you should NOT use GenAI

As artificial intelligence becomes rapidly integrated into our everyday lives and workplaces, it's easy to get swept up in the promises of efficiency and automation.

But in our community of Quality Engineers and Testers, we will not blindly embrace Generative AI without considering the risks, won’t we?

In this talk, Rik, based on his extensive experience of both Quality Engineering and Testing on the one hand and Generative AI on the other hand, will take a hard look at the reality of using GenAI and explore why, in quite some cases, you should NOT use it.

Let’s delve into the significant risks that come with relying on GenAI. From the presence of bias in AI models, to over-automation, which risks taking the human touch out of decision-making.

We'll also discuss the troubling phenomenon of AI "hallucinations," and the many privacy and security concerns that remain a major issue, as GenAI often requires access to sensitive data that is not always properly safeguarded.

And Rik will elaborate about people themselves being a risk, because they don’t know how to properly create prompts to apply GenAI in a safe and useful way.

Yet, this talk isn’t just about avoiding GenAI. Rik will also explore the useful applications of GenAI and how you can use it wisely and safe.

Situations where GenAI truly shines are augmenting decision-making processes, handling text-heavy tasks (including various kinds of reviews) and processing vast amounts of data in ways that would otherwise be impossible for humans to manage.

In conclusion we will demonstrate, with practical examples, that to fully harness its potential, quality engineers need more than just technical knowledge. Skills like critical thinking, exploration, and curiosity become essential when working with GenAI.

Join me for a balanced yet thought-provoking discussion about the future of GenAI in Quality Engineering, where we'll examine both the limitations and opportunities AI brings to our profession.