Sarus is a privacy layer that sits between data sources and analysts, enabling data querying and analysis without direct data exposure. It intercepts queries submitted by authorized analysts, rewrites them with privacy guarantees enforced on the fly, and returns results while keeping the underlying data protected. Deployment: Sarus can be deployed on-premises or in public cloud environments (AWS, Azure, GCP) using Docker or Kubernetes. Setup is designed to be completed in minutes. How it works: - Administrators connect Sarus to one or more data sources and define which tables are available for querying. - Access rights are granted to specific data consumers, with configurable privacy policy settings per user or query. - Analysts submit queries using familiar tools such as SQL, pandas, or scikit-learn. Sarus automatically rewrites the query to enforce privacy guarantees while maximizing result accuracy. Key capabilities include data connectors for linking to various data sources, configurable privacy policies, audit logging, synthetic data generation, support for analytics and BI workflows, machine learning integrations, and the ability to act upon data. The product is aimed at organizations that need to enable analytics and AI workloads on sensitive data while complying with privacy requirements. Analysts never have direct access to the raw data; all interactions are mediated through the Sarus privacy layer.