SPONSOR TRACK TALK

Quality Engineering for the Age of AI: How to Test Learning Systems

The emergence of artificial intelligence—particularly models that learn and evolve—has radically transformed the quality of software. This session puts the evolution of testing in the context of AI and highlights the importance of adapting quality engineering practices to ensure trust, security and value in AI-powered products.

In this talk, Javier López Perdiguero will discuss how to apply Quality Engineering to AI-based systems. He will analyse the challenges of testing generative AI, LLM-based solutions and autonomous agents, stemming from their non-deterministic and data-dependent nature. He will present a validation framework covering functionality, security, bias, robustness and explainability, alongside continuous evaluation practices integrated into LLMOps pipelines. In addition, AI-specific quality metrics, techniques for evaluating hallucinations, prompt security and the use of datasets and synthetic data will be explored.

What you’ll learn

Session details

Javier López Perdiguero

Javier began his professional career 25 years ago as an SAP consultant specialising in financial modules, and after 10 years he began to transition into the world of software quality as a Test Lead, where he has served as a testing project manager. For the past few years, his role as Delivery Manager at Sogeti has focused on ensuring the successful and timely delivery of high-quality software products, combining team management with responsibility for final delivery. What he likes most about QA is that his projects go into production without defects, minimising negative impacts on his clients, with the resulting cost savings that this entails.