We are an early-stage deep-tech company solving one of the hardest problems in AI deployment: making models run reliably on constrained edge hardware, across any silicon.
The edge AI market is growing fast, but deployment remains a pain point. Teams spend months porting models to specific silicon, only to repeat the work for the next chip. Nuatara's engine abstracts this complexity — one integration point, all your hardware targets.
Founded in 2026, we combine deep expertise in compiler backends, hardware abstraction, and silicon bring-up to build infrastructure that others can't easily replicate.
Our team comes from embedded systems, imaging, and AI runtime engineering. We work close to the metal — driver layers, firmware, and the toolchains that sit between a model file and a chip.
We validate everything on actual target silicon. Benchmarks on a laptop mean nothing for a vision chip on a factory floor.
We work on the compiler and runtime layer — not model training, not dashboards. Deep, unglamorous infrastructure that makes everything else possible.
We hire slowly and carefully. Everyone here owns their domain completely and works on things that genuinely matter.
We do not have formal job listings — we hire for exceptional people who want to work deep in the stack. If you are excited about compiler backends, embedded systems, hardware bring-up, or AI runtime engineering, we want to hear from you.
Tell us what you would work on. We are a small team building hard infrastructure — the right people make all the difference. Reach out directly and let's talk.
Talk to us about your target silicon, your model, and your deployment constraints.
Talk to us →