Friday 13th June 2025 - 11am, Kiril Gashteovski (NEC Laboratories Europe)

Speaker: Kiril Gashteovski (NEC Laboratories Europe)

Title: Grounded Intelligence: Towards Reliable and Explainable LLM Systems through Synthetic Data, Evaluation, and Modular Design

Abstract: Large Language Models (LLMs) are increasingly used in many real-world scenarios. Their usability, however, is dependent on more than task performance: it also depends on our ability to build reliable and explainable LLM systems, which are grounded in evidence. We approach this goal by resting on three pillars: synthetic data generation, evaluation strategies, and modular system design. In this talk, we are going to discuss each of these pillars in details, thus shedding light on our vision for building reliable and explainable LLM systems.