With the widespread adoption and deployment of machine learning across enterprises, ever-increasing amounts of data are being collected, stored, communicated, combined, and computed over by sophisticated algorithms. In parallel, new governmental regulations and rising concerns about privacy are giving impetus to new research on how to protect the confidentiality, integrity and privacy of all this data.
At Cryspen, we are engaged in multiple R&D projects around these topics. We help design new encryption standards, such as HPKE, for protecting data at rest; we implement high-assurance cryptographic libraries, such as libcrux; we design and implement new protocols, such as MLS, for communicating sensitive data between a large number of endpoints; and we investigate state-of-the-art constructions, such as multi-party computation, for privacy-preserving analysis over distributed data.
We are working with the SINE Foundation on the Atlas project, whose goal is to build a privacy-preserving data custodian for municipal data. In particular, we are helping to transition research results from the academic partners, Anja Lehmann at HPI and Florian Tschorsch from TU Berlin, by building usable specifications as well as the cryptographic components needed for the data custodian. In this project, we will use our hax toolchain, the hacspec language, and our verified crypto library libcrux. We will also help design and analyze the multi-party computation components needed in Atlas.
Cryspen is also an associated partner at PRAIRIE, the Paris Artificial Intelligence Research Institute. PRAIRIE aims to bring together academic research groups in the area of machine learning with industrial partners, like Cryspen, who can help deploy research results. In particular, Cryspen aims to provide high-assurance privacy-preserving data analysis components and cryptographic libraries to interested partners within PRAIRIE.
Eager to explore the world of privacy-preserving data analysis or delve into other intriguing topics? Reach out