Celato is building a federated inference layer for real-world evidence, enabling pharmaceutical companies to run research queries across distributed hospital data without moving patient records. The platform uses agentic study design and a versioned query language (MDQL) to deliver exact, auditable answers while keeping data under hospital control. It accelerates RWE generation by up to 10x while eliminating data-movement risks.
Funding
Funding not disclosed
Founders
Product
Problem
Real-world evidence generation is slowed by fragmented, siloed patient data that cannot be moved or centralized due to privacy, governance, and logistical constraints. This forces pharma and regulators to rely on incomplete study populations, delayed insights, and costly, slow data-access processes that weaken medical decision-making.
Solution
Celato provides a federated inference layer that lets pharmaceutical companies run research queries across distributed hospital data without any patient-record movement. The platform's agentic layer translates natural-language research questions into precise, versioned study designs and MDQL definitions, which are then executed locally at each hospital node under strict governance. Results from all nodes are aggregated into a single, mathematically exact answer with a full audit trail, while scientists remain in the loop at every step. This approach delivers insights up to 10x faster while preserving 100% hospital control and zero risk to patient data.
Target Audience
Primary customers are pharmaceutical companies, clinical research organizations, and regulatory bodies that need faster, more complete real-world evidence from distributed patient data while maintaining strict data governance.
Features
- Agentic study design that decomposes natural-language research questions into precise, executable study protocols
- MDQL (Medical Data Query Language) for versioned, reproducible definition of study cohorts and endpoints
- Federated execution engine that runs queries locally at each hospital node, ensuring data never leaves the source
- Source-level harmonization of schemas, codes, and free text across heterogeneous hospital systems
- Mathematically exact aggregation of node-level results into a single auditable answer with full traceability
- End-to-end audit trail from question to output, supporting regulatory-grade evidence generation