ML / AI roles at companies known for safe to fail. Companies where it's genuinely safe to take risks, make mistakes, and speak up. Blameless post-mortems, experimentation culture, and supportive management.