Briefing: Machine learning arrives in practice: An engineer's field guide to a shifting technical landscape

This briefing equips practising engineers with the conceptual vocabulary needed to evaluate machine learning tools now entering routine use. In particular, predictive models are distinguished from generative systems, and both families are mapped onto the software an engineer is likely to encounter. The discussion is grounded in three case studies – a predictive model that evaluates fire resistance, a generative large language model that iteratively derives empirical indices from experimental records and a computer vision model that classifies damage from imagery. The briefing then turns to the questions a practitioner should keep asking at the forefront of data provenance, uncertainty and accountability, and closes with a glossary that distinguishes technical meaning from popular assumptions across ten terms now used in the profession daily.