John Cummings, PhD

I trained and deployed learning solutions for Sonar, MTech’s IoT SaaS product, providing insights to farmers on thousands of farms at a given moment, each containing tens of thousands of chickens. While at MTech, I ● Developed machine learning models using Python and libraries such as XGBoost, Scikit-learn, Pandas, SQL, and more to analyze sensor data, providing inference and predictive models for our clients. ● Deployed my machine learning models that introduced new features for our product and improved upon existing features using scalable cloud solutions written using Python querying our SQL/NoSQL databases. ● Analyzed data, found trends, and maintained communication with company stakeholders regarding the development of our product using these data to address our business needs and strategies. ● Conducted ad-hoc analyses for various value-generating reports such as analysis of strategies optimizing flock performance and swiftly identifying sensor hardware issues. ● Communicated results and insights to non-technical audiences using analysis tools such as Matplotlib/Plotly visualizations, SHAP game-theoretic feature analysis, partial dependency plots, and causal inference.

As a tenure-track assistant professor of mathematics at UNG, I ● Won the UNG Presidential Summer Incentive Award where I created numerical solvers for fluid flow problems in C++ using the deal.II finite element library. ● Won an MAA Project NExT Fellowship in which I adopted and improved classroom methodologies resulting in outstanding teaching as evidenced by my students’ evaluations and peer observation. ● Collaborated with colleagues on mathematical research into numerical solvers for partial differential equations modeling fluid flow, managed research projects for upper-level mathematics students, and completed professional development programs to improve my teaching and research. ● Implemented effective pedagogical techniques to teach four classes per semester such as statistics, ordinary differential equations, numerical analysis, calculus, and more to approximately 80 students per semester. ● Served on numerous committees to complete services needed by the university such as hiring new faculty, fostering community within the department, and contributing to the effective operation of our annual mathematics competition, among other services.

During this one-year visiting appointment as an assistant professor of mathematics, I ● Published an interdisciplinary paper on the modeling of organic photovoltaics using efficient numerical solvers written in Fortran 90. ● Mentored a group of students to a second-place finish in the local chapter of the national ordinary differential equations competition SIMIODE hosted by SCUDEM. ● Taught four classes a semester including ordinary differential equations, business calculus, and numerical analysis to 80 students per semester

Taught two mathematics classes ranging from precalculus to numerical analysis to about 40 students per semester while I made progress on my doctorate in the form of successfully completing classes, passing qualifying examinations, and writing my dissertation. Dissertation: Mathematical Modeling of Mixtures and Numerical Solution with Applications to Polymer Physics