Computationally proficient research mathematician with expertise in Python, supervised deep learning, unsupervised deep learning for dimensionality reduction and latent space discovery, PyTorch machine learning frameworks, bash scripting, topology-informed regularization, and high performance computing. Highly adept at clustering and classification. I focus on synthesizing machine learning tools to find intimate trends in real-world phenomena, including biological and x-ray scattering images. My work is readily integrated with other fields, as evidenced by my collaborations with biologists, physicists, theorists, and experimentalists in other domains. Current work: leveraging sparsity to alleviate big-data problems. I augment existing convolutional neural network (CNN) ideas to create new architectures with many fewer (stochastically-connected) nodes/layers. This allows for the training of multiple leaner networks in parallel, which may then be aggregated or used for uncertainty quantification.Products: lead developer of DLSIA (Deep Learning for Scientific Image Analysis, formerly pyMSDtorch), a comprehensive PyTorch-based deep learning library for image analysis tasks, such as pixel-by-pixel segmentation, denoising, inpainting, and fast outlier detection (sub 1 ms for 64^ pixels). With a focus on user-customizability and a friendly API, DLSIA offers fast instantiation of Autoencoders, Tunable U-Nets, Mixed-Scale Dense Nets, and more.More generally: I work with many teams building end-to-end frameworks and solutions for a variety of machine learning tasks for image data, including x-ray scattering data (FXS, SAXS/WAXS, etc.) and biological image data from a variety of microscopy modalities (lattice light-sheet, FIB-SEM, cryo).Research interests include: - neutral collapse in applications to Security and machine UNlearning,- self-supervised ML for automated labeling, - CNN-derived latent/feature space exploration,- sparse deep learning architectures, - vision transformers, - sparse-sampling quadrature design.