Carolyn Smith Email and Phone Number

Trained a differentiable surrogate model for optical signal transport in neutrino experiments using PyTorch Lightning deep learning framework. Validated model performance through interactive data visualization and statistical methods.Developed Python package allowing users to train and use differentiable model for a variety of detector analyses.

Implemented a machine learning model utilizing SciPy and scikit-learn Python libraries to predict reservoir volumes based on satellite-derived areas. Building a data pipeline to integrate NASA's satellite radar data with visualization software to enable real-time monitoring and display of global reservoir volumes.

Developed and optimized a noise reduction pipeline for inertial measurement unit data collected from patients with osteoarthritis. Analyzed activity and joint loading patterns from this data.