• Built hierarchical mutli-label text classification application in Python using Pytorch to classify products scraped from the web into Amazon hierarchies. Put together training/testing pipeline. Model trained on 4,434,908 unique Amazon training points representing 15,520 labels. Model tested on 31,369 unique non-Amazon data points and achieves P@1 of 97% and P@3 of 91%. Produced two deployment versions. First version is a dockerized Python/Flask/Bjoern microservice. Second pulls records off of RabbitMQ queue and runs on both CPU and GPU. Achieves labeling rate of about 200 records/second on an NVIDIA Tesla V100 GPU. Gained more than four times speedup by running data preprocessing (tokenization) and postprocessing (reconstructing hierarchies from ordered labels) as parallel python processes. Application exhibits horizontal scaling across machines.• Performed and presented data analysis on distribution and recency of business, product, and property data in production Elasticsearch.• Maintained and re-wrote legacy Java libraries, including use of JUnit testing with Mockito.• Helped test and maintain nodejs applications.

I helped developed linear and integer programs for General Motors's Supply Chain Purchasing and Planning divisions. I modeled the linear and integer programs in Java, Python, and the Optimization Programming Language (OPL) and OPL Script (a derivative of Javascript) built around IBM's CPLEX solver. I did data analytics with Python, including some use of the Pandas library.

I tutored all the levels of calculus, real analysis, linear algebra, differential equations, matrix analysis, physics, and statistics. I also tutored Python.

I helped teach Integral Calculus for Science Majors.

I helped teach Integral Calculus for Science Majors.