Software engineer / Personal projects
I’m Jaidon.
I build what
interests me.
Learning German. Running AI locally. Turning images into dots. This is a collection of the things I’m making and the ideas I keep coming back to.
Projects & experiments / 01—04
Different interests.
A common thread.
Each of these started with something I wanted to learn, use, or understand. Some are apps. Some are experiments. All of them give me a reason to keep building.
01 / iOS · Language learning
KarteiDeck
Three languages.
One little habit.
Learning German started it. KarteiDeck now brings German, Russian and French into one free flashcard app, with 3,000 words per language, spaced reviews and separate progress for every course.
Swift / SwiftUI / iOS
Visit KarteiDeck ↗Inside KarteiDeck
Start with 5, 10 or 15 new words, then return for spaced reviews. Russian has its own alphabet practice: meet letters, listen, draw from memory and recall them over several days. Pronunciation guides are optional. No account, subscription or AI coach is required. The three-language update is submitted to Apple and awaiting review.
02 / Desktop · Development prototype
Kane AI
Local intelligence.
Visible work.
I’m interested in what local AI can do when it has useful tools and I can see its work. Kane is my desktop harness for exploring that: local models, explicit permissions, and a visible record of what happened.
Rust / Tauri 2 / Ollama
Visit Kane ↗Inside Kane
Kane uses Rust, Tauri and Ollama. It connects local models to development tools, with role-based permissions and recorded tool outcomes. It’s a supervised prototype I’m continuing to explore.
03 / Local learning · Research prototype
Trifallax
Start small.
Study what stays.
Trifallax is where I explore learning on a smaller scale. It connects a small recurrent network to a local console so I can train it, save its state, and study what it recalls.
Python / MLX / GRU
Visit Trifallax ↗Inside Trifallax
This Python and MLX experiment trains on generated patterns and submitted messages. Saved training state and bounded recall checks let me investigate retention. It’s an early learning experiment; broader language ability is still an open question.
04 / Creative tools · Early prototype
Pointloom
Light into points.
Motion into texture.
The artwork on this site became a project of its own. Pointloom turns images and video into grayscale dot art—an experiment in depth, texture, and finding a different way to see familiar things.
Python / Depth inference / Local rendering
Visit Pointloom ↗Inside Pointloom
Choose a local file, infer its surface depth, and export a PNG or silent looping MP4. The current prototype runs locally; its automatic background removal handles plain white backgrounds. The site’s portrait uses a separately refined recording pipeline.
Beyond the project names
A place for what
I’m into.
I’m a software engineer, and this is a space for the things I choose to build outside my day-to-day work.
Language learning, local AI, and visual experiments keep pulling me in different directions. I like being able to follow an idea far enough to make something I can use, look at, or learn from.
Sometimes one project leads to another. The dot artwork here started as a way to make this site feel like me. Working on it led to Pointloom.
- Learning German
- Local AI
- Creative code
- Small experiments
Good ideas start with a conversation
If something here
speaks to you, say hello.
I’d like to hear what you’re making, learning, or exploring—especially if our interests overlap.
jaidonclinton@icloud.com