Research
Focused labs. Serious problems.
Our research is organized into specialized labs, each owning a domain where AI can create lasting value — with its own agenda, teams, and products.
Surgical intelligence
Microvision: vision models for the operating room
Microvision applies computer vision to surgical scenes — instrument tracking, phase recognition, and scene understanding — toward safer, better-instrumented procedures across African health systems.
StatusActive research
FocusSurgical intelligence
DomainHealthcare · Computer vision
Language / multimodal imagery
Multimodality
Frontier capabilities for low-resource languages
Multilinguality, post-training, and vision-language models for the languages the field leaves behind — making frontier capabilities work for African speakers first.
StatusActive research
FocusMultilinguality · Post-training · VLMs
DomainLanguage technology
Publications
Research, in the open.
We publish our work where possible and share it with the research community. Publications will be listed here as they are released.
Collaborate on research →