Biography
I am a PhD candidate applying machine learning and artificial intelligence (AI) to public safety. Most of my published work revolves around forecasting --- using machine learning to predict who will and who will not commit a crime in a way that is both ethical and statistically sound. My models can typically predict crime more effectively than the status quo of professional judgement or unvalidated risk assessments. As such, I believe they can be used to foster better criminal justice decisions --- decisions that result in fewer low-risk individuals incarcerated, a reduction in serious crime, and are based on instruments that have partially remediated historical biases commonly found within criminal justice data.
My other projects involve using language models to sift through large amounts of police data, as well as use computer vision to detect violence in both body-worn video and CCTV footage.
I have sat on a steering group for the Home Office, as well as lectured for the National Crime Agency, Home Office, and University of Cambridge. I have also worked with police departments across the USA and UK, as well as taught courses on big data, data visualisations, and AI as it relates to policing.
Prior to my PhD, I completed my undergrad at Harvard University, where I graduated magna cum laude with highest honors. I have also completed a visiting studentship with the University of Oxford and obtained an MPhil degree from the University of Cambridge in Criminological Research.
See https://jverrey.com/ for my personal website, which was designed for industry.
Research
Member of the Jerry Lee Centre of Experimental Criminology.
Publications
Verrey, J., Ariel, B., Harinam, V., & Dillon, L. (2023). Using machine learning to forecast domestic homicide via police data and super learning. Scientific reports, 13(1), 22932.