Amine LargouAmineLG
For companies and individuals: Python services and internal tools, AI features that stay under your control, and the Kubernetes and OpenShift platforms they run on. AmineLG is my software business. I do the technical work myself, I also join teams on contract, and I build my own tools.
What I do
For small businesses, product teams and individuals. Every project starts with a small first step you can say yes to.
Someone spends hours every week copying data, rebuilding reports or babysitting a script.
You want AI in a real workflow, but it can’t be allowed to guess, invent data or act without approval.
Releases are stressful, nobody is sure who can access what, and outages are noticed by customers first.
Software your business depends on keeps breaking, and the person who wrote it has moved on.
My own software
Alongside client work I build my own tools, mostly in the open. They share one idea: a model can make useful judgments, while ordinary, tested software controls validation, permissions and execution.
2026 · Open source · public alpha
Ask a SQLite database a question in plain English, see every decision behind the answer, and approve each change before it happens.
The clearest example of the approach: the model picks from options the code has already checked, every decision is visible, and nothing is written without confirmation.
2024 · Open source · 129 stars on GitHub
Turns folders of plain notes into structured Obsidian Markdown with a model running on your own computer. With the default setup, nothing is sent anywhere else.
Where my AI work started: the model does one narrow job where the data already lives, and the software around it stays plain and predictable.
Read the case study Source on GitHub (opens in a new tab)Video demo (opens in a new tab)
A playground for comparing AI decision models on the same input, and seeing what changes their answer.
3 models compared on one input
Explains who can access what in an OpenShift cluster, and why, including access that quietly comes back.
480 tests · early v1
A live real-time game, deployed from Git to my own Kubernetes cluster with TLS, monitoring and hardened defaults.
Live a.aminelg.com
All my work: client projects, own software and upstream fixes
How this direction took shape
Local AI models first, then the infrastructure to run software reliably, then a clearer approach that ties the two together.
2024
I started building seriously with local AI models. Obsidian Text Transformer was the first concrete project.
2025–26
I went deeper into Kubernetes, OpenShift and the infrastructure that keeps software running reliably.
2026
IntentSQL and Decision Lab: models make useful judgments, while ordinary software controls validation, permissions and execution.
Today
I use all of it to build software for clients, and to develop my own tools through AmineLG.
The foundation
Since 2017 I’ve built and supported software for paying clients in the US, Australia, New Zealand and Singapore, mostly trading systems and the services around them, running unattended in live markets. That includes two working relationships of eight and nine years. It’s where my production habits come from.
Amine has consistently demonstrated high technical logic and provided the reliable uptime required for all of our business operations.
…his ability to clearly understand the requirements of what have been at times involved and complex projects, and his skills to deliver the coded project on time and to above the agreed standard.
From written recommendations, January 2026, about trading software I built and supported. Shared with permission, initials only. The client work, in detail
Describe it in your own words; no specification needed. I read every message myself and reply within two working days, often sooner.
or email contact@aminelg.com