Data Scientist

• Jul 2022 - Present

Optimization Researcher

• Jan 2021 - Present

‐ Ran experiments in MATLAB to determine the relationship between entropy and heterogeneity, in order to solve the seeded graph matching problem

Machine Learning Researcher

• Apr 2020 - Dec 2020

‐ Built a neural network class from scratch using PyTorch to quickly process hundreds of thousands of data points ‐ Researched the effect of a neural network’s architecture on its performance, finding that networks with more sparsely connected nodes run faster and produce well‐fitting models ‐ Analyzed and fine‐tuned the network using Monte Carlo simulations and hyper parameter adjustments

Consulting Intern

• Jun 2020 - Present

‐ Analyzed sentiment of thousands of tweets, revealing a correlation between surprise, disgust, and false news ‐ Built a sentiment‐based natural language processing model using BERT for a tool which halts the spread of online misinformation ‐ Organized and advised a NATO war game to determine the top technical solution to widespread misinformation out of 47 companies ‐ Published a paper in the NATO review and presented to the Czech Ministry of Defense ‐ Analyzed sentiment of thousands of tweets, revealing a correlation between surprise, disgust, and false news ‐ Built a sentiment‐based natural language processing model using BERT for a tool which halts the spread of online misinformation ‐ Organized and advised a NATO war game to determine the top technical solution to widespread misinformation out of 47 companies ‐ Published a paper in the NATO review and presented to the Czech Ministry of Defense

Data Scientist

• Feb 2021 - Nov 2021

‐ Extracted features for time‐series data using Numpy and Pandas and built more robust models by selecting top features using Sci‐Kit learn, A/B testing, and Pearson correlations ‐ Improved missile tracking and false threat detection using computer vision analysis and machine learning, increasing classification model accuracy by 95% ‐ Programmed Monte Carlo iterations in parallel and built visual plots, like ROC curves, in MATLAB to optimize clustering thresholds ‐ Collaborated with… Show more ‐ Extracted features for time‐series data using Numpy and Pandas and built more robust models by selecting top features using Sci‐Kit learn, A/B testing, and Pearson correlations ‐ Improved missile tracking and false threat detection using computer vision analysis and machine learning, increasing classification model accuracy by 95% ‐ Programmed Monte Carlo iterations in parallel and built visual plots, like ROC curves, in MATLAB to optimize clustering thresholds ‐ Collaborated with Networked Sensors and Integrated Fires team in the Air and Missile Defense department ‐ Obtained Active Secret Security Clearance Show less ‐ Extracted features for time‐series data using Numpy and Pandas and built more robust models by selecting top features using Sci‐Kit learn, A/B testing, and Pearson correlations ‐ Improved missile tracking and false threat detection using computer vision analysis and machine learning, increasing classification model accuracy by 95% ‐ Programmed Monte Carlo iterations in parallel and built visual plots, like ROC curves, in MATLAB to optimize clustering thresholds ‐ Collaborated with… Show more ‐ Extracted features for time‐series data using Numpy and Pandas and built more robust models by selecting top features using Sci‐Kit learn, A/B testing, and Pearson correlations ‐ Improved missile tracking and false threat detection using computer vision analysis and machine learning, increasing classification model accuracy by 95% ‐ Programmed Monte Carlo iterations in parallel and built visual plots, like ROC curves, in MATLAB to optimize clustering thresholds ‐ Collaborated with Networked Sensors and Integrated Fires team in the Air and Missile Defense department ‐ Obtained Active Secret Security Clearance Show less

IBM Watson Research Intern

• 2018 - Apr 2019

‐ Used machine learning classification and fMRI data on ~80 subjects to predict conversion to psychosis in schizophrenic patients ‐ Published paper in the SPIE Medical Imaging Journal ‐ Won the Acorda Scientific Excellence Award, presented research on the Lisa Wexler Show, and won 4th Place in Bioinformatics category at WESEF, a top science fair ‐ Used machine learning classification and fMRI data on ~80 subjects to predict conversion to psychosis in schizophrenic patients ‐ Published paper in the SPIE Medical Imaging Journal ‐ Won the Acorda Scientific Excellence Award, presented research on the Lisa Wexler Show, and won 4th Place in Bioinformatics category at WESEF, a top science fair