Software Engineer developing simulation models for robotic solutions.

Adapted Fully Hyperbolic Neural Networks (FHNN) to embed functional brain connectivity graphs, derived from magnetoencephalography (MEG) data, into low dimensional Lorentzian hyperbolic space. Developed hyperbolic embedding pipeline to detect changes in intrinsic hierarchy of functional subnetworks as well as predict age for patients across multiple decades from the Cam-CAN dataset with multimodal MEG and MRI data.

Developed CLI to automate simulation software code pipelines and to improve user experience.− Automated installation process for simulation software, making onboarding process simple and scalable, as well as making process more user-friendly by requiring less background knowledge.− CLI capable of running simulations, visualizing simulation runs, gathering key simulation metrics, and incorporating new software through CLI commands, making handling of simulation software efficient.

Analyzed machine learning pipelines for time series anomaly detection. − Examined reconstruction accuracy vs. anomaly detection precision in models via F1 and MSE metrics.− Evaluated prediction-based models employing LSTM networks and reconstruction-based models using Auto-Encoders and GAN networks.− Preprocessing, modeling, and post-processing univariate and multivariate time series

Developed software framework for the Plasma Science and Fusion Center (PSFC) to model three-dimensional laser experiments on OMEGA laser in Rochester, NY based on configuration inputs.