Coming soon
The benchmark for AI that makes human decisions.
We’re building the evaluation environment for high-stakes human decisions, starting with hiring. Real assessments, real outcomes, and the reward signals that teach models to judge like the best humans do.
Building or training in this space? Reach out and we’ll talk.
The Environment
The most realistic simulation of high-stakes human evaluation, captured end to end.
The Benchmark
Measurable, outcome-linked tasks that score how well any model judges real candidates.
The Reward Model
Trajectories tied to real outcomes. The exact signal labs need to train better judgment.
