This role represents a new specialization in AI safety infrastructure—specifically building tooling and data systems for frontier model safety evaluation and training. This category has emerged only in the last 2-3 years as enterprises scale AI deployment and regulatory pressure around model safety intensifies.