We're building sovereign AI that runs on your hardware, not someone else's cloud. Small team, enormous ambition, zero compromise.
We deploy to production every day. No month-long sprints. No death by committee. Build it, test it, ship it.
Every byte stays on our hardware. Zero cloud dependency for inference. Privacy isn't a feature — it's the foundation.
We train models from scratch, build our own infrastructure, and question every assumption. No black boxes.
We run 22+ models on local hardware. No API bills. Constraints breed creativity — we find the trick.
Train and optimize sub-3B parameter models for on-device inference. Experience with PyTorch, model quantization (GGUF/GGML), and Metal/CUDA required.
Build and maintain distributed inference across heterogeneous hardware (Mac Studio, MacBook, Raspberry Pi). Networking, PM2, nginx, and llama.cpp experience preferred.
Red-team our models and infrastructure. Gradient leakage attacks, membership inference, prompt injection — break things so we can build them stronger.
Next.js 14, React, TypeScript. Build the control plane for sovereign AI — dashboards, chat interfaces, training monitors, deployment pipelines.
Work from anywhere. We're distributed across Canada.
Mac Studio, GPUs, whatever you need to build.
Your work ships publicly. Build your portfolio while you work.
No layers. Direct impact. Every commit matters.
Conferences, papers, courses — stay sharp.
Ship daily, but on your schedule. Results over attendance.
We're always looking for exceptional people. If you're obsessed with sovereign AI and want to build the impossible, reach out.
GET IN TOUCH