CryptoNumerics offers a comprehensive product suite to help companies break through their Data Barrier to generate Data-Driven Insights — from organizations that simply want to protect the privacy in their data for internal analysis, to organizations that want to build statistical and machine learning models using decentralized and cryptographically protected data.

Our products are designed to work in concert as an integrated data engineering and machine learning pipeline to generate Data-Driven Insights.

Secure Identification of common elements in private datasets


Eliminate the need for “trusted” third parties by securely identifying common elements across sensitive datasets directly

CN-Fuse uses Private Set Intersection to securely identify agreed upon common elements without revealing anything else and without the exorbitant cost, time, and risk of using a “trusted” third party.


  • Securely identify common data elements across n parties

  • Runs locally in each party’s environment with secure peer-to-peer communications

  • Prepare datasets for decentralized secure machine learning (CN-Insight)


Protect all elements that could cause privacy leakage

Privacy concerns block your ability to analyze and work with data.

CN-Shroud allows you to publish a privacy-protected dataset where the risk of re-identification has been balanced with data quality through tunable parameters.


  • Protect personal identifiers, quasi-identifiers, and sensitive attributes

  • Couple differential privacy with anonymization to create a privacy-protected dataset

  • Choose between a variety of metrics that best define data quality for your application

  • Choose from k-anonymization, l-diversity, t-closeness and others to find the technique that produces the optimal dataset for your application

Decentralized Secure Machine Learning


Build machine learning models using decentralized privacy-protected datasets without moving them

Data owners are unwilling to negotiate access to their proprietary data with you, because they fear loss of IP or because regulatory constraints prevent its relocation.

CN-Insight allows you and your data partners, both internal and external, to train statistical and machine learning models on datasets without the need to centralize or relocate them. This is accomplished by a combination of cryptographic protocols, Secure Multiparty Computation, and numerical methods resulting in intellectual property protection.



  • Specify Model ownership among n parties

  • Federated learning across n parties

  • Concurrent learning across n parties

  • Runs locally in each party’s environment with secure peer-to-peer communications