Marc Warfield — AI Safety Practitioner
I can run a model safety evaluation end to end — threat-model a deployment, author the policy, red-team it responsibly, measure harm and over-refusal, reason about alignment and interpretability limits, manage a risk register against a recognized framework, and make a defensible go/no-go recommendation — with an artifact behind every one of those verbs.
One deployment, walked from worry to sign-off — the continuous story behind the binder. Each link is backed by an artifact below.