Currently exploring ai interpretability and app layers