· Led development of SCoPe machine learning classification project/repository, achieving up to 98% precision/recall· Trained CNN/XGBoost algorithms, using Weights & Biases to track experiment performance, perform hyperparameter sweeps· Identified most important features based on XGBoost training, decreasing feature computation requirements for future studies· Optimized Python feature generation/inference algorithms to run on billions of time series, conserving limited HPC resources· Implemented new API/frontend features for Python/ReactJS/Postgres web app in response to user feedback, scientific goals· Developed NMMA Bayesian analysis repository, used ML to improve model grid interpolation speed from hours to minutes· Built Python API connecting HPC-supported analysis service to web app, enabling new automated workflows· Maintained automated documentation, added new unit tests, prepared repositories for PyPI/conda release· Hosted weekly interdisciplinary team meetings, collaborated with scientists and engineers to achieve research goals· Mentored 4 undergraduate / 6 graduate students, provided constructive code reviews, supported completion of mentee goals

· Designed custom SQL queries to obtain archival data from multiple relational databases· Built Python workflows to run statistical analyses of real and Monte Carlo simulated data· Ran parallelized Python code on Amazon Web Services EC2 to expedite analysis workflows· Published results in peer-reviewed scientific journals, wrote funding proposals (awarded $58,752 grant)· Hosted observatory events, led physics demonstrations for visiting members of the public

· For Senior Thesis, revised scientific literature by using Python to wrangle data and run MCMC analysis· Recognized as top departmental researcher in graduating class

· Used command-line interface on Linux operating system to analyze solar images

· Used GIS to map, calculate ages of Martian lava flows· Performed public education/outreach on museum floor