Put your intelligence to work.
We build in public, publish our research, and credit our contributors. The work is foundational and the team is small. Both are intentional.
Your work will outlive the product.
Language infrastructure compounds. Every model we ship, every corpus we validate, every community dataset we open-source becomes a permanent part of the NLP landscape for these language families. The impact timeline is measured in decades.
The hardest NLP problems are here.
Low-resource modeling, morphologically rich languages, offline edge inference, tonal acoustic modeling, consent-based data collection at scale. If you need a challenging problem, we have a surplus of them. No shortage of frontier work.
The team speaks the languages.
Most of our engineers, linguists, and researchers are native speakers of the languages we model. That is not a coincidence. Cultural and linguistic proximity to the problem produces better science. We hire accordingly.
6 roles
Don't see your role listed?
We hire for demonstrated capability, not job titles. If your skills are rare and your values match ours, reach out with what you can do — not what you have done before.
Write to us directly →