Sarus

What is Sarus?

Sarus is a privacy-preserving data analytics and synthetic data generation platform designed to enable analysts and data scientists to query and manipulate sensitive datasets without direct access to the source data. The software utilizes differential privacy and artificial intelligence (AI) to synthesize data, obfuscate individual records in aggregate results, and enforce governance policies across analytical workloads.

Key Capabilities

Native Data Integration: Connects directly to relational databases, data warehouses, and cloud storage systems (such as Amazon Simple Storage Service (S3) or Google Cloud Storage (GCS)) to process data without requiring extraction or replication.
Differential Privacy Enforcement: Applies mathematical noise to query outputs through centralized privacy policies, ensuring that individual data points cannot be reconstructed from analytical results or machine learning (ML) models.
High-Fidelity Synthetic Data: Automatically generates structural replicas of original datasets—preserving multivariate correlations across relational tables, numerical data, and text—for use in exploratory analysis and testing.
Remote Execution Capabilities: Provides a Python Software Development Kit (SDK) and Structured Query Language (SQL) Application Programming Interface (API) that rewrite and execute code (using libraries like pandas and scikit-learn) on remote data, returning only policy-compliant outputs.

Audience & Use Cases

Audience: Data scientists, data analysts, data privacy officers, and machine learning engineers.
Use Case: Secure machine learning model training, cross-site data analysis, analytical querying via Business Intelligence (BI) tools, and sharing safe data inside secure clean rooms.

Technical Specifications

Supported Integrations: PostgreSQL, MySQL, Azure SQL, Snowflake, Databricks, BigQuery, Amazon Redshift.
Supported Interfaces: Python SDK (pandas, NumPy, scikit-learn), SQL API, HiveServer2 for BI connections (Tableau, PowerBI).