Developing user-focused software for model analysis under uncertainty
Accelerating materials development with data science
Naive model metrics can be misleading when using AI for materials discovery. I introduced techniques to better assess a model's suitability for guiding materials development.
Peforming model analysis under uncertainty is uncommon and challenging. I developed Grama to make uncertainty quantification easier to teach and to communicate.
Present aircraft sizing criteria can be inefficient or unsafe. In this work I develop provably conservative alternatives.
High-dimensional simulation data can be difficult to analyze. In this work I merged physical insights with data-driven techniques to perform *data-driven dimensional analysis*.