Summary
Andrew Demetriou is a research scientist blending nine years of interdisciplinary experience across industry, non-profits, and academia to redesign how we collect, validate, and present data for machine learning. Currently finishing a PhD at TU Delft, his work pioneers psychometrics-informed ground-truthing for subjective, high-variability tasks—demonstrated in case studies on song lyrics and political speech and by leading the Alexandria federated micro-publication project. He has translated that research mindset into product-impacting roles at Universal Music Group and VIER, building lightweight distributed ML evaluations, multi-sensor data pipelines, and gamified simulation environments. Known for bridging social science rigor with practical engineering, he distills complex experimental results into actionable insights for stakeholders and prototypes. Based in Amsterdam, he combines deep methodological critique (including on LLM overinterpretation) with hands-on tooling and visualization to make subjective data auditable and useful.
9 years of coding experience
11 years of employment as a software developer
Bachelors Degree Political Science; Philosophy, Bachelors Degree Political Science; Philosophy at CUNY QUEENS COLLEGE
Master’s Degree Social Psychology Research, Master’s Degree Social Psychology Research at Vrije Universiteit Amsterdam (VU Amsterdam)
English, Greek, Dutch