Karl Penzhorn

For my PhD, I am building a finite element gradient plasticity code in C, tying together various systems I wrote myself—an LU matrix solver, a matrix library with preallocation, an iterative Newton-Raphson solver for the non-linear equations, cross-platform stack traces, graphical animations of the output, and multi-threading. Wrote high-performance code for GPUs using CUDA and using multithreading locally. Also converted Matlab to C using outputs to compare results one line at a time.

Helped design algorithms and implemented the code for predicting the population of fish based on various data. Created 1D and 2D smoothers to allow for taking derivatives on the data, which we had planned to input into a Newton-Rhapson type algorithm to find optimal parameters in our model. Also wrote various solvers to optimize for speed as well as using GPUs for parallel processing through the nVidia library CUDA, which we ran on Google's Cloud infrastructure.

Upgraded our in-house ticketing system to allocate remote jobs, which were scheduled in a calendar and manageable through our administrative interface. It also exposed a REST API so our mobile app could connect to it, allowing contractors to view their schedules and locate and update job statuses, all through their phones. It also allowed admin staff to be notified when jobs escalated for being late or overdue.

Migrated a platform from an existing Oracle-based ERP system to new MySQL database tables, mapping fields, validating data, such as customer names and addresses, and generating reports for offending results. Customers, contracts, products, and the entire financial systems were all done through independent modules using a common code.

Won the bid among our top developers to retool our front-end interface platform by designing a modular JSP-based component and layout system.

Many of our clients had their own in-house developers who needed to be trained on the new system they were going to support. I was tasked to develop the courses and exams for training and certifying external coders with our system architecture, development tools, and best practices.

Recently, one of our larger clients wanted to start charging for interest on overdue accounts, for some 350,000 customers. Along with another developer, I wrote the system that enabled the charge calculations based on timestamped financial transactions in the financials core. To understand how much each customer was overdue at any point in time, we took things like dynamic prime lending rates and custom charge dates into account.

Wrote tutorials for a SaaS company on how to set up and create React and React Native applications that are authenticated and that access and control a remote data display.

Using various Python libraries (e.g., Facebook Prophet) I implemented a REST API for a startup company to process time series data coming from Asana, producing things such as decomposition and forecasting.