Pushing the boundaries of privacy-first AI and applied natural science.
Four interconnected areas of research driving our product development and scientific contribution.
Practical machine learning systems for real-world deployment. Focus on reliability, interpretability, and production-readiness.
Computational approaches to natural science problems. Bridging traditional research with modern AI capabilities.
Building systems where privacy isn’t a feature—it’s the architecture. Zero-trust, zero-knowledge, full sovereignty.
AI that works without the cloud. Optimized models for resource-constrained devices and disconnected environments.
Challenges we’re actively working on—and open to collaborating on.
Building effective clinical NLP models for Indian languages with limited training data.
Processing video streams for engagement detection without any identifiable information leaving the device.
Training models across hospitals with vastly different data formats and infrastructure.
Interested in Collaborating?
We’re always looking for researchers, institutions, and organizations to work with on these challenges.
Get in TouchSelected research outputs and technical reports.
Privacy-preserving synthetic data generation for healthcare ML
On-device clinical NLP for low-resource languages
Differential privacy in classroom analytics
Zero-knowledge document sharing protocols
Edge inference for real-time engagement detection
Federated learning for multi-site healthcare studies