I have worked on numerous projects throughout my career as both a programmer and environmental scientist! My passion primarily lies in the intersection between data, algorithms, and how they can support the clean energy transition, but I have loved working on projects in a wide variety of areas.
Python, R, MATLAB
As part of an internship with Florida Atlantic University in 2021, I worked with a research team within the Institute for Sensing and Embedded Network Systems Engineering (I-SENSE). The research primarily centered on the optimization of buoyancy-controlled ocean curren turbines through the application of control co-design (CCD) methodologies. Optimization methods often utilized dynamic & linear programming methods. Computationally-expensive aspects of the design were optimized through supervised machine learning processes. Results are currently under review for appearance in various conferences & will (hopefully) eventually be accepted for journal publishing.
Python (Jupyter Notebook)
Transition-Metal Complexes (TMCs) are molecules consisting of one or more transition-metal centers surrounded by organic molecules (known as ligands), bound to the metal atom by dative bonds. These molecules are of interest to researchers because of their potential to mimic the efficiency and specificity of catalytic biological molecules, such as enzymes. However, engineering these molecules for use in catalysis is difficult and computationally expensive, owing to the large combinatorial space of of transition metals and ligands. As part of our final project for our machine learning course, myself and another student used a simple machine learning approach to predict the metal-center charges of TMCS based off structural features of the molecules. Additionally, these methods were theorized with the purpose of being abstract enough to be easily expanded to other electronic properties in order to eventually facilitate the prediction of complete electronic profiles of TMCs. Machine learning has emerged in the chemical sciences as a method for producing models of complicated systems with large amounts of variables. While these methods have been extensively applied to areas such as drug discovery, there have been little applications to the discovery and modeling of TMCs. Here, we use simple machine learning methods to predict the metal-center charges of TMCs based off structural features of the molecules. The use of these methods could be easily expanded to other electronic properties with the goal of creating models capable of predicting complete electronic profiles of TMCs.
As part of an internship with the National Renewable Energy Laboratory in 2020, I worked with a team of other developers on augmentation and testing of REopt Lite (now known as REopt Web Tool). Throughout this process, I developed Python and R algorithms to collect and analyze data on the geographic viability of natural gas power across the United States, presenting the results in the form of a paper & poster at a conference. Upon completion of my research & testing, I wrote & published API documentation for REopt Lite.
As part of the Computer Science + Social Good organization at UNC-Chapel Hill, I worked on a team of students to help develop a web application for the campus food pantry. We primarily worked with the goal of aiding users and volunteers with privacy and ease-of-use, utilizing a simple relational database to facilitate inventory bookmarking and allow users to "hold and "request" items.
Kay Blada Recycling
As part of the Computer Science + Social Good organization at UNC-Chapel Hill, I worked on a team of students to support a Haitian recycling organization. We provided data analysis and visualization to advise them on future business practices, as well as redesigning their website and identifying key datapoints for them to collect in the future.