EZKL is a cryptography research and development company focused on proactive security for AI/ML systems. We use zero-knowledge proofs to help prove that models produced certain results and verify computation, without reviewing the model itself. At EZKL, we believe in open-source first and build our core engine in public. This enables new affordances in trust, whether you're seeking to monitor in-system models, verify model performance publicly, or identify model-generated content. We're on a mission to create robust standards in AI safety.