Research lab that produces training data for frontier AI labs to handle ambiguous, non-deterministic, long-horizon professional tasks.
Traverse partners with frontier AI labs to produce training data that enables models to develop taste and judgment for subjective and long-horizon tasks in domains like law, healthcare, writing, and strategy. The lab captures real experts operating in real environments and preserves their reasoning process, creating training signals that teach models how judgment actually works. Its long-term goal is to give frontier models the foundations needed to perform and eventually surpass human white-collar work. Traverse is a product of Traverse.