Summary
Roman Marshanski is an AI generalist and creative technologist with a decade of experience applying human-centered methods to train and refine generative models, currently shaping RLHF workflows at Outlier AI. He pairs deep practical experience in content creation, SEO, and WordPress site-building—having grown Humoropedia.com from zero to thousands of monthly visitors—with hands-on cybersecurity analysis of dark web threats and botnet research. Comfortable switching between rigorous evaluation tasks and creative writing, Roman brings a rare blend of editorial judgment, threat intelligence rigor, and prompt-engineering craft to model alignment. Based in Los Angeles and grounded by formal study in computer information systems, he thrives at the intersection of creativity and technical problem-solving, often turning nuanced human feedback into measurable ML improvements.
10 years of coding experience