I support the advanced analytics function of the human resources group, as we leverage the largest set of hospital and healthcare data in the world. My current focus is on RN retention. We are using AutoML models developed in partnership with Google in conjunction with a Streamlit app developed in house to infer the causes of turnover and retention. Previous work involved an application designed to forecast vacancies and net headcount gains. The application involved selecting a best-performing candidate time series model to predict nurse terminations. I worked to improve the predictive accuracy of the forecasts, and to lead its migration from Microsoft Azure to the Google Cloud Platform.

Creator of the RichardOnData YouTube channel, a fast growing channel with 24,000+ subscribers and over 850,000+ views. The objective of the channel is to provide the audience with perspectives with overviews and tutorials of different programming languages (R, Python, Julia), coverage of statistical methodologies ranging from regression methods to survival analysis, overviews of hot topics in the data science field at large, and guidance to newcomers on how to break into data science and start careers of their own.

My work everyday centered around a high throughput real-time decision system. The underlying data platform combined Hadoop and Spark to leverage enormous amounts of data. My work was in the Ad Optimization group - a team whose role it was to optimize every lever of this system. One key project involved building an ML model that determined whether, given advertisements that met various business KPIs throughout their flights, whether they would be likely to do so if more impressions of them were served. Another project involved exploratory analysis of datasets with tens of millions of rows, to determine why certain line items had unusually low start rates.

I worked with a leading multi-national pharmaceutical organization to provide support to various analytical efforts. These included: applications developed in RShiny aimed at providing quick insights to medical teams and biostatisticians, and machine learning endeavors aimed at generating inference about and predicting safety and efficacy of new drugs. Later, I worked to provide internal support towards the development of new data science processes and platforms, built a full-stack introductory data science course offering, and led a dashboard development project for a small pharmaceutical client.

I worked in teams -- often in agile product development environments -- to provide modeling and analysis for healthcare performance improvement projects for a variety of clients. Software used included SQL, R, Shiny, and Excel. One endeavor involved implementing a methodology to classify wasted pharmacy dispense events with a most probable root cause. Other projects included staffing and occupancy analyses, predictive model development of sepsis in an inpatient environment, prototyping various reports and decision support software using Shiny, and discrete event simulation to determine birth volume capacity based on facility constraints.

I taught two lab sections of Statistics 250: Introduction to Statistics and Data Analysis. Responsibilities included teaching lab, grading homework and exams, holding office hours, and proctoring exams. For one year I was also a mentor for new Graduate Student Instructors. Over the course of five semesters, I had anywhere from 54 to 117 students per semester.

My responsibilities included making a plan to switch Matthaei Botanical Gardens and Nichols Arboretum's electronic database from Microsoft Access to KE Electronic Museum, correcting errors within the institution's database, and various research projects. I performed statistical analysis for a research project which is published in the National Academy of Science in Belarus.

I helped teach a two week long summer camp course called Life, Death & Change: Landscapes and Human Impact. This was through the Michigan Math and Science Scholars program, which is an internationally recognized program for high school students. I led discussions on statistical techniques and on how to approach data. I helped teach this course three times.

I provided consulting services to analyze various academic data. Several regression and machine learning techniques were used. All models were built using the statistical package R.