What's under the hood?
Federated learning
Send models to the data instead of data to the models. Using federated learning, Shoji lets you analyse multiple datasets at once as if they were in the same place.
Secure computation
Share the use of data, not the data itself. We use SMPC, a cryptographic technique that allows several parties to jointly compute results from their combined data without seeing or storing your data.
Differential privacy
Keep your customers protected every step of the way. Shoji uses a mathematical definition of privacy to guarantee that outputs never reveal any personal information.
