• Processed voltage data through Artificial Neural Network, using Linux based workload manager SLURM, to model activation of nerve fibers near Deep Brain Stimulus electrodes with 99% validation accuracy• Generalized model from one to all possible electrode configurations by increasing the amount of data 4100%• Cut data generation time by 90% through python script, eliminating the need to launch a Finite Element Modeling application and transfer files• Improved model… Show more • Processed voltage data through Artificial Neural Network, using Linux based workload manager SLURM, to model activation of nerve fibers near Deep Brain Stimulus electrodes with 99% validation accuracy• Generalized model from one to all possible electrode configurations by increasing the amount of data 4100%• Cut data generation time by 90% through python script, eliminating the need to launch a Finite Element Modeling application and transfer files• Improved model performance 15% after identifying optimal hyperparameters of an exhaustive search, spanning the number of hidden layers and epochs, batch size, and eight other hyperparameters, on TensorBoard Show less

• Matched patient to best financing option with a 0.63 silhouette score, employing K-means as the machine learning algorithm for a multivariate dataset absent of target values• Streamlined data preprocessing pipeline, automating comma separated value data to pandas data frame• Increased model applicability on real world data via Mahalanobis distance metric, relaxing Gaussian assumptions for K-means• Eliminated 10 dimensions, while explaining explaining 90% of variance… Show more • Matched patient to best financing option with a 0.63 silhouette score, employing K-means as the machine learning algorithm for a multivariate dataset absent of target values• Streamlined data preprocessing pipeline, automating comma separated value data to pandas data frame• Increased model applicability on real world data via Mahalanobis distance metric, relaxing Gaussian assumptions for K-means• Eliminated 10 dimensions, while explaining explaining 90% of variance using Principal Component Analysis through Sklearn, allowing for better data visualization Show less