Summary
Mark Kurzeja is a Staff Research Engineer at Google DeepMind with a decade of experience building and securing large-scale machine learning systems and applied Bayesian models. He has progressed from designing deployed Bayesian defenses for billions of Android devices to leading frontier-model alignment and post-training mitigation efforts at DeepMind. His background blends rigorous applied statistics (Master’s in Applied Statistics) with finance and math roots, having modeled multi-hundred-million-dollar CLOs earlier in his career. Mark brings practical production expertise across foundation model training, sparse ML, ranking, and combinatorial algorithms, and he teaches complex statistics topics to large audiences—evidenced by a Coursera course with 80,000+ enrollees. Colleagues know him for a thoughtful, skeptical approach to certainty—reflected in his GitHub bio—and for translating probabilistic reasoning into robust, production-ready AI safety systems.
10 years of coding experience
6 years of employment as a software developer
Master's degree Applied Statistics, Master's degree Applied Statistics at University of Michigan - Rackham Graduate School
Minor in Mathematics and Minor in Computer Science, Minor in Mathematics and Minor in Computer Science at University of Michigan
Bachelor of Business Administration (B.B.A.) Finance, Bachelor of Business Administration (B.B.A.) Finance at University of Michigan - Stephen M. Ross School of Business