
Club 5060 Ti
Practical local LLM presets for RTX 5060 Ti systems, each backed by the benchmark receipt that proves it.
$ whoami
I'm Spooky, a self-taught developer in Queensland. I build local AI tooling, web products and the occasional strange experiment, and I run the Linux infrastructure they live on.
Most of it started as something I needed or wanted to understand. Each project here says plainly where it stands: live, maintained, experimental or paused.
open source, products and client work

Practical local LLM presets for RTX 5060 Ti systems, each backed by the benchmark receipt that proves it.

An open, inspectable laboratory for embodied connectome intelligence: stimulus, spiking network and movement in a 3D terrarium.

A local-first cycle-tracking PWA. Works offline, needs no account, and keeps data on the device.

A guarded Linux kit for running one GGUF model across AMD Vulkan and NVIDIA CUDA cards, powered by llama.cpp.

A photo-led website and identity for a family cleaning business in Yeppoon.

A self-hostable workbench for species distribution modelling, developed with Unstable Branch.
Route-planning and operations software for depot and field work, shaped around the sequence of a real working day.
Exclusive GPU leases for cooperating local inference engines.
A research-first Agent Skill for writing rigorous, self-contained prompts that another agent can execute without follow-up.
An interactive explainer: two sensor arrangements with the same power spectrum and very different ways to fail.
Benchmarks and local LLM recipes for 16 GB AMD Radeon GPUs.
public GitHub, forks excluded
4,331 contributions in the last year · 248 active days · snapshot 5 Oct 2026
languages across 17 own public repos
Websites, internal tools, or a problem that needs a closer look. A short note is enough to start.