ORIOLETECH
Curiosity,
made tangible.
Exploring local intelligence, practical software, and the questions worth following.
Explore the work01Parallel Latent ReasonerExperimental AI
An experimental MLX prototype exploring recurrent latent deliberation with pretrained Gemma models on Apple Silicon. The repository records the evidence and limits.
View research ↗02System 1Local decisions
Teaches bounded decisions from examples, checks a candidate before adoption, and runs the saved skill locally. It is not an LLM.
View research ↗03LatticeShadowWork memory
A macOS work-memory prototype with explicit capture choices and scoped assistant access. Its documentation distinguishes tested behavior from open validation work.
View research ↗04τ-GateDependency research
An experimental Rust tool for analyzing dependency graphs. It explores a security problem; it does not guarantee that a dependency or system is safe.
View research ↗25 years in technology.
Still asking questions.
I'm Stephan Arrington. My recent work centers on local LLMs, experimental AI, and practical software. OrioleTech is where that curiosity meets focused research and engineering.
What are you exploring?
Research collaborations and paid consulting inquiries are welcome. Tell me about the problem you want to solve.
Discuss a project