Hi, I'm Paul (^-^)/
I am an electrical engineer interested in topics where software interacts with the physical world. So far this has mostly focused around various topics related to electric vehicles. I currently work at DeepDrive GmbH where I work on a variety of software topics related to the development of motor inverters for novel automotive drive units.
Focused on inverter software, end-of-line testing, and vehicle control unit development. Developed a patent-pending rotor position sensor and built full-stack Python tooling for automated hardware testing and database management.
Converted the PyTorch implementation of Centermask to Recogni chip math, maintaining a relative performance drop of just 0.25% on the Coco 2017 dataset, and contributed to toolchain debugging.
Developed and deployed a full-stack web application (React, Flask) on an internal Kubernetes cluster to monitor manufacturing data, complete with a Jenkins CI/CD pipeline and comprehensive automated testing.
A university project expanding on HyperDiffusion to generate neural fields using diffusion models for novel 3D shape generation, featuring a pre-processing pipeline for PartNet meshes.
A web application built around the NeuralProphet library. I designed a new front-end in Figma, gathered user feedback, and implemented the final design in React to facilitate an improved user journey.
Designed the electronics and embedded software for an emergency braking system, implemented long-range in-car radio telemetry, and actively commissioned the vehicle across test sessions and track days.