AI safety researcher focused on red teaming and agentic AI, translating research into practical solutions for robust, aligned systems. Publications, experience, and writing on interpretability and alignment.