NLP Meetup

Doors open around 7pm for drinks and pizza, talks kick off at 7.30pm and from 9pm there’s time to exchange stories, ideas, and the latest AI gossip.

We’ll start the evening with Michael Bauwens and Heike Pauli from the GPT Academy at UCLL, who will demonstrate how companies can evaluate LLM applications in a rigorous and practical way. The two other speakers will be announced in the next few weeks.

Making LLM evaluation tangible for non-technical professionals
Michael Bauwens and Heike Pauli (UCLL | GPT Academy)
Many SMEs adopt generative AI tools out of fear of missing out, often without a structured evaluation of whether these open-ended, LLM-based systems truly meet their needs. Initial “vibe checks” or ad hoc testing frequently lead to premature judgments, overlooking how performance depends on task, context, and usage. This talk presents a work-in-progress research project that develops an accessible, holistic evaluation framework for non-technical end-users in Flemish SMEs, enabling them to systematically assess GenAI systems against concrete tasks, expectations, and example outputs across multiple dimensions. Drawing on a pilot study with a GenAI platform for social workers and experiments with LLM-as-a-judge methods such as G-Eval, Michael and Heike share practical insights, demonstrate an evaluation platform, and discuss limitations, showing how structured evaluation can help SMEs make more informed decisions, optimize GenAI investments, and build sustainable confidence in these technologies.

Want to register? Click here