Retrieval-Augmented Generation (RAG): advances, challenges and business insights

Data Science Platform Seminar Series VI

Speaker: Karl Cini

Talk Abstract

Retrieval-Augmented Generation (RAG) combines large language models with external retrieval mechanisms to ground generated outputs in relevant, up-to-date sources, improving factual accuracy, currency, and explainability. This talk traces the technical origins of RAG and its core architectural building blocks, and surveys current production-grade implementations, including dense, sparse, and hybrid retrieval approaches, vector databases, and reranking strategies.

It critically examines the practical limitations and risks of RAG systems, such as retrieval errors, latency, cost, and governance concerns, and explores how the emergence of very large LLM context windows is reshaping the design tradeoffs between retrieval-based and context-heavy approaches. The session concludes with a pragmatic business decision framework, addressing cost, compliance, and performance metrics, to help organizations determine when and how RAG should be adopted within enterprise AI systems. Concrete evaluation techniques and deployment considerations are discussed throughout.

This is the sixth talk in the 2025-2026 Data Science Platform Seminar Series.

Speaker’s Bio

Karl Cini is a Research Support Officer within the Department of Artificial Intelligence at the Faculty of ICT, University of Malta, and a Fellow of the Association of Chartered Certified Accountants (FCCA) with over 25 years of experience at the intersection of data, finance, and business strategy. He holds a Master of Science in Data Science from the University of Malta, where he developed advanced expertise in AI integration, data analytics, and data modeling to support data-driven decision-making.

Prior to specialising in data science, he built a strong commercial and advisory career as an accountant and business consultant, working with both local and international clients. Alongside his academic role, Karl advises organisations on the practical and responsible adoption of AI and data-driven solutions, supporting strategy definition, analytics-led decision making, and organisational readiness for digital transformation.

Add this event to your calendar.