My research is driven by one principle: security guarantees should survive deployment. I design cryptographic protocols for distributed systems whose guarantees hold when the environment departs from the traditional models they were proved in, spanning secure multi-party computation, Byzantine consensus and the cryptographic infrastructure of blockchains, from impossibility results to protocols that run in practice.
Most secure protocols assume a known bound on message delay that no real network can promise. We removed that dependence from two sides. A series of works brought asynchronous MPC, which makes no timing assumption at all, down to linear communication with optimal resilience, settling a question open since the early 1990s, and then rebuilt it on hash functions and symmetric primitives so that it runs among hundreds of parties (Velox). Network-agnostic security, which we introduced for MPC, gives one protocol the optimal guarantees of a synchronous protocol if the network happens to be synchronous and those of an asynchronous one otherwise; our latest results show this robustness comes essentially for free.
Key papers: Velox (CCS 2025) | Computationally Efficient Asynchronous MPC (CRYPTO 2025) | Towards Achieving Asynchronous MPC with Linear Communication (CRYPTO 2024) | Perfectly Secure Network-Agnostic MPC Comes for Free (EUROCRYPT 2026) | Always Have a Backup Plan (CRYPTO 2020)
Consensus is the primitive under all distributed security, and two costs decide whether it runs at scale: communication and latency. We gave the first fully asynchronous Byzantine agreement with subquadratic communication, matching lower bounds, leader election at polylogarithmic cost per party, and agreement on long messages at essentially the cost of sending them once. On latency, a paradigm we introduced led to the first protocols meeting the 1986 Karlin-Yao round bound, open for over thirty years. Current work turns both costs into quantities that track the network and adversary actually present: early termination, round-optimality without trusted setup, and responsiveness without sacrificing validity.
Key papers: Round-Optimal Byzantine Agreement (EUROCRYPT 2022) | Asynchronous BA with Subquadratic Communication (TCC 2020) | Leader Election with Poly-logarithmic Communication (CRYPTO 2025) | Communication Lower Bounds for Cryptographic Broadcast (DISC 2023, Distributed Computing 2025) | Round-Optimal BA without Trusted Setup (EUROCRYPT 2026)
Long-lived infrastructure changes: machines are replaced, organisations join and leave, adversaries accumulate footholds. Our work on MPC with dynamic committees lets small, changing subsets of participants carry out different parts of a computation so that no adversary ever controls a whole committee at once. We gave the first unconditionally secure protocols in this model, settled when agreement is feasible with dynamic committees, and wrote the unifying survey of the area. This line is the subject of our SNSF project (2026-2030).
Key papers: Perfect MPC over Layered Graphs (CRYPTO 2023) | Statistical Layered MPC (TCC 2024) | Feasibility of Broadcast with Dynamic Committees (CRYPTO 2026) | Efficient Distributed Randomness Generation (EUROCRYPT 2025) | A Unified View of MPC with Dynamic Committees (JIIS 2026)
The group's research is supported by the Swiss National Science Foundation (project grant, 2026-2030), the Canton of Zug through the Zug Blockchain Initiative, Parity Technologies, the Stellar Development Foundation, the Ethereum Foundation, Protocol Labs, the Hasler Foundation, the Swiss Quantum Initiative, and ETH Zurich's Leading House Asia.
Recent Advances in Asynchronous Secure Computation. Simons Institute Workshop on Secure Computation, Berkeley, July 2025.
Multi-Party Computation with Dynamic Committees. TPMPC, IISc Bangalore, March 2025.
Network-Agnostic Security: Importance and Challenges. Dagstuhl Seminar 24362, September 2024.
Towards Achieving Asynchronous MPC with Linear Communication. Bertinoro Workshop on Cryptography in the Blockchain Era, June 2024.