• Developed and maintained a solution to mock production formatted simulation data for rating and underwriting impact analysis to convert ensure accuracy during legacy conversion. ◦ Used a Jupyter-Notebook configured to run a Pyspark instance to take multi-dimensional SAS and SQL sourced dataframes and join them together via a custom defined python UDF, which repackages and formats a proprietary json / edn payload, mocking the format needed by the models. ◦ Debugged and wrote tooling to identify causes of simulation errors, using regex, Pyspark, and Python. • Rewrote simulation preparation for operational glue job to new core format and simultaneously increased performance by 35% by using spark functions conjoined with Python list comprehension.• Wrote custom flattening AWS glue job that takes a proprietary hierarchical json simulation response format and flattens it to a singular data frame for analytic consumption by actuaries. ◦ Generalized the script to be capable of flattening to any target dimension, including risk, policy, driver, etc. ◦ Optimized performance to cost less than $5 per 1.8 million record execution ◦ Documented using Google’s Python style guide.• Maintained simulation-data-pipeline, a Terraform managed GitLab repository containing information on our established model monitoring systems, including step functions which operate on specially queried Redshift jobs which are fed into a step function containing a simulation on EMR and several subsequent glue jobs.• Trained formally in Data Scientist lead course on GLM development, using scikit-learn for modeling and seaborn for visualization techniques.• Worked with legacy DB2 based databases to extract data using SQL for use in impact analysis.• Enforced higher documentation standard by assisting in identifying needed retro topics for team self reflection as part of the agile life cycle.

• Worked with Spark to run simulations on AWS EMR cluster to validate accuracy of thousands of attributes• Developed python based Jupyter notebook to reverse engineer flattened data into hierarchical .json files for testing data flattening process.• Updated and operated with PySpark notebook to convert flattened .csv-like data into .edn compliant hierarchical schema defined data using Pandas for the purposes of running model validation through a simulation pipeline.

• Developed tools and improved existing infrastructure used by GLMs for both underwriting and actuarial consumption.• Wrote sweet recursive functions in Clojure to optimize work flow, improve performance, or to just do something cool.• Acquired insurance business acumen needed to understand the process flow for customers.• Utilized bash scripts to optimize workflow and improve effeciency. • Familiarized myself with data sources available to actuarial modeling and pricing.• Practiced best developer techniques including version management using git, gitlab, github and AWS CodeCommit.• Endured agile methodology for the purpose of ensuring transparency, organization and best developer practice.

Handled receiving packages, asset management, and filing paperwork for the universityFacilitated transactions totaling over $200,000 worth of technology productsEducated 500+ students regarding campus technology during scheduled presentationsAssisted in training new employees and developing training documentsSolved customer service issues

Assisted faculty with day-to-day classroom operations including proctoring examinationsAllocated time outside of scheduled work periods for grading and other department workTutored students during both office hours and scheduled tutoring hoursConducted classroom lab sessions alone weeklyAccurately maintained student grades