Hi, I am Paul. Thank you for visiting my portfolio! I am a physics graduate from the University of Cambridge. Currently, I work as a Python software developer for coppafisher. I love problem solving, software development, applied mathematics, and game development.
Coppafisher is an open source software used to analyse microscopic images, usually of mice brain tissue. Coppafisher aims to correctly read hundreds of unique gene types found in the brain in real space, near the forefront of spatial neuro-transcriptomics. The software tackles many challenges like image filtering, registration, gene calling, large data, code collaboration, testing, stability, and optimisation. I have worked with many Python packages, like Google's new machine learning toolbox jax, numpy, pandas, napari, and matplotlib.
I love open source software and I use it daily! What passionate people do as hobbies inspires me. I particularly enjoy using the Neovim text editor. I have worked on my own plugins for Neovim, like a paragraph formatter called textangle.nvim.
I work on a roguelike, 2D strategy game in Godot. My growing experience in software development and collaboration improved my skills in writing maintainable, readable, and scaleable code. This allows me to make my game a reality that I first imagined as a child.
Iceland is geologically active. Thousands of micro-earthquakes can occur everyday. The Cambridge Volcano Seismology Group (CVSG) has been recording Icelandic seismicity for over a decade. Over the years, a lot of the data has been picked for seismic events by human analysts. This is ideal data for a neural network to mimic the job of a seismologist. I tested and retrained a neural network known as PhaseNet to pick the arrival times of micro-earthquakes in Iceland. It was my first time working with machine learning package tensorflow and getting hands-on experience with open source code.
