Systems software developer working primarily on Linux, with a focus on the layer between the kernel and the desktop: distribution engineering, service management, boot-time infrastructure, and the small utilities layered on top of it.
Most projects begin at a gap: a tool that does not exist, or one considerably heavier than the work it performs. The method that follows is usually the same — take the layer apart far enough that rebuilding it becomes the only way to be certain it is understood.
The repository list is best read as a record of systems I wanted to understand from the inside rather than as a product portfolio. One constraint applies throughout:
It has to work on my own machine first — minimal dependencies, and nothing included on the chance it might be useful later.
- Linux internals & distribution engineering — init and service management, package tooling, boot chain, and filesystem layout.
- Automation & desktop tooling — written for portability and minimal runtime dependencies.
- Applied machine learning on biological data — connectome simulation on the FlyWire Drosophila dataset, and reward-driven control of embodied agents in place of hand-authored behaviour scripts.
Arch Linux and BlackArch as primary environments. Hand-written systemd units and package definitions, bootloader and initramfs configuration for custom images, and QEMU used to validate builds before they reach physical hardware.