Submission for G20 TechSprint 2025
Consumer-consented and secure credit data portability
Read the detailed problem statement here
Privacy-Preserving SME Rating Engine
Vist the deployed platform here
The Privacy-Preserving SME Rating Engine allows financial institutions to assess SME creditworthiness without ever moving or exposing raw data. Built on a Multi-Party Computation (MPC)-based Cryptographic Computing Virtual Machine (CCVM), the solution ensures that SME financial information remains encrypted at source, while lenders only receive verified inferences such as repayment capacity or credit risk. Each computation is cryptographically bound to the SME’s consent, ensuring strict adherence to approved usage and regulatory compliance, including data localisation requirements.
This model reduces the trust gap between institutions & unlocks access to previously inaccessible data in a secure way. Instead of relying on centralised data sharing or third-party intermediaries, banks can collaborate to gain only the information they need for underwriting. This improves model precision, supports financial inclusion by serving viable but underserved SMEs, and ensures compliance and privacy are built into every step.
By using cryptography as the trust layer, the system creates interoperability across open banking ecosystems and can support a wide range of collaborative use cases. Whether for cross-border SME lending, credit history portability, or broader open finance applications, it enables institutions to work together without exposing raw data, overcoming regulatory and operational barriers while improving efficiency, accuracy, and trust.
Follow this quick setup guide, SETUP.md, to run the application locally.
See LICENSE file for details.