Worked with team members to build out a new robotic process automation service using UIPath RPA solution. Part of team that produced first live process in EPIC.Currently working on moving a no-show- prediction model into a Pilot phase with estimated returns of over 1M annually once it is live.

Built a model to predict likelihood of a patient not showing up for a scheduled appointment with estimated hospital gains reaching millions of dollars. Initial analysis suggests an AUC of 90 percent with recall over 50%. Recently the model was tested for performance if extended to a new area of patients. The existing model still performed at 80% AUC and a quick retrain brought performances close to 85%. As a team of four, rebuilt an existing staffing model for improved speed (minutes down to under 1 minute) and improved algorithm and UI (fewer clicks, improved and corrected logic, including generation of better probability distribution curves). This has led to interest from outside organizations to commercialize the product. Tools used include Python, Django, SQL, GitHub, JavaScript, and Java and .Net based programs in Anylogic and UIPath robotic process automation.Random Forest, Neural Networks, Logistic Regression, Optimization, data extraction, forecasting, principal component analysis PCA.

I took a project analyzing EEG time series data from processing to data visualization to feature extraction and then model building and validation. My analysis has led to a submitted research paper dealing a discovery that what people thought was random noise might in fact contain significant information about pathological tissue. Tools include machine learning algorithms such as neural networks and logistic regression, principal component analysis (PCA), signal processing, cross validation, and model valuation. Data visualization includes using scatter plots, heat maps, ROC curves, and histograms to find trends, correct problems with the data, and understand relationships between features. Submitted large memory (60-120 GB) jobs to the UofM cluster for analysis using PBS script files. Matlab scripting, Linux environment, neural networks, logistic regression, parallel computing toolbox, cross validation, time series data, etc.I have also learned the basics of brain functionality, high frequency oscillations, and seizures.

Research project on nuclear physics resulted in improved understanding of the proton by analysing nonlinear data using a modified gradient descent algorithm (Levenburg Marquardt). I overcome significant problems such as sparse data, bias vs variance, and a large 600+ parameter space to improve understanding of the proton. I learned to use plots and numbers to tell a story to my advisor on why changes to the model were essential leading to a successful project. During my PhD work, I modified both analysis codes to fully automate the analysis steps, speeding up the analysis significantly. I also updated the codes from Fortran 70 to Fortran 95.I also created a shell script to allow base parallelization to be incorporated into the fitting routine to speed up the analysis. Added additional features to the code to improve understanding of how changing parameters affected the fits.

Laboratory assistant for introductory physics labs. Managed two lab groups each semester. Answered students questions, and gave physical insights into the labs to aid in student comprehension of physical concepts. Graded homeworks and lab assignments.

Created and presented an hour long planetarium show to school aged children between the ages of 6-18. The first 30 minutes was a star show discussing the stars and constellations, their history, and their importance. The last thirty minutes was a powerpoint show discussing modern advances in knowledge, current scientific missions, and visual photographs of the frontiers of astronomy including from the Hubble Space Telescope. Following the presentation was a brief Q&A session for the children to ask questions and make comments. I especially enjoyed linking the stars and constellations in a way that brought them to life to the students such as using mythology, seafaring stories, and pop culture.