Summary
Ben Rachbach is a Machine Learning and Evaluation Product Manager with a decade of experience blending software engineering, research, and startup-building to improve how AI answers research questions. At Elicit he leads ML roadmap and a dedicated evaluation team that creates automated, high-fidelity evals—work that drove a ~78% precision improvement in search and underpins the product’s core data-extraction feature. He also designed evals and tasks for published research on iterated decomposition and has hands-on experience translating customer needs into ML specs and MVPs. Earlier roles range from technical PM and experimenter at Ought to leading engineering teams at Wonder Workshop, demonstrating a rare mix of product rigor, empirical research practice, and shipping production systems. Based in the Bay Area, he combines human-centered design instincts with a quantitative, experiment-driven approach to make AI outputs more accurate and vettable.
10 years of coding experience