A security platform focused on protecting machine-learning models and AI pipelines from adversarial attacks, model theft and supply-chain risks in the model files and training pipeline themselves.