Edwige Cyffers develops mathematical tools and safeguards to ensure that machine learning is trustworthy and to help end users make informed decisions about their participation. Her primary area of interest is privacy: she uses differential privacy to protect data in machine learning systems, particularly in decentralized contexts where participants retain control over their data. More broadly, her research examines the long-term effects and societal impact of the deployment of machine learning.
Edwige Cyffers is a researcher at the CNRS (research fellow) at Paris Dauphine-PSL University. In the spring of 2026, she was a Research Fellow in the “Federated and Collaborative Learning” program at the Simons Institute for the Theory of Computing. Previously, she was a postdoctoral researcher at the Institute of Science and Technology Austria (ISTA). Her doctoral dissertation, supervised by Aurélien Bellet at Inria Lille, received the L’Oréal-UNESCO For Women in Science – Young Talents Award and the SSFAM Award, and was a finalist for the AFIA Dissertation Award.
She also promotes women in science, particularly mathematics, and recently co-organized the European Girls’ Mathematical Olympiad in 2026. She is committed to decentralizing machine learning conferences and co-organized NeurIPS in Paris.
Edwige Cyffers is the recipient of the Peccot Lecture for the 2026–2027 academic year, upon the recommendation of Professor Timothy Gowers, holder of the Combinatorics Chair.