Saif Malik

I work on governance-first AI — architectures that decide not just what to predict, but how to behave when evidence becomes uncertain, contradictory, or corrupted. The goal is failure-mode control: systems that fail safely under stress.