Sarus

Sarus provides a privacy layer for analytics and AI, enabling data scientists to work with sensitive data while enforcing privacy protections. The platform combines differential privacy, privacy-preserving synthetic data, and on-the-fly query rewriting to keep data secure during analytics and LLM workflows.

Sarus sits atop data sources, allowing data practitioners to query sensitive data without direct access. It uses differential privacy to guarantee safe query results, generates privacy-preserving synthetic data, and rewrites processing jobs on-the-fly to enforce privacy. It supports deploying on-premises or in the cloud (Docker/Kubernetes), connects to data sources, and controls query access and privacy levels for each user. Outputs are guaranteed to comply with privacy policies, enabling use of SQL, pandas, sklearn, and LLMs with private data handling.

Who it’s for: Organizations handling sensitive data across industries such as finance, healthcare, marketing, and public sector, with data science and analytics teams requiring privacy-safe analytics and AI workflows.

Funding mentions (EIC Funding, ELSA Innovation Lab) and Azure Confidential Clean Rooms integration indicate active traction and collaboration with research and cloud ecosystems.