Summary
Ben Smith is an ML engineer and former industry data scientist turned postdoctoral researcher who blends neuroscience-grade fMRI analysis with practical AI safety and human evaluation work. With 11 years of experience spanning Apple, Facebook, academic labs, and startups, he designs rigorous human-centered evaluation pipelines, autograders, and multi-objective models that compare AI agents directly to expert human performance. His work bridges neural models of value and decision-making (including large-scale fMRI and behavioral datasets) with product-facing safety frameworks and benchmarks used in real launches. Notably, he has operationalized end-to-end evaluation systems for engineering agents and led stakeholder-aligned safety assessments at scale. Based in the Greater Seattle Area, he combines deep statistical rigor and experimental design with hands-on implementation of ML evaluation tooling and mentorship in AI safety.
11 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (PhD), Psychology, Doctor of Philosophy (PhD), Psychology at University of Southern California
Pukekohe High School
Master of Science (MSc), Psychology, Master of Science (MSc), Psychology at University of Auckland
English, Chinese, Chinese