Fan Zhang

Recent Projects

A unifying theme of my research is building trustworthy computer systems through three approaches: (1) decentralization, which enables programs to run on consensus networks without a trusted operator; (2) confidential and verifiable computation, which uses cryptographic proofs or trusted hardware to let users verify correct execution or protect private inputs; and (3) accountability, which gives users evidence of misbehavior. We develop foundational building blocks with the goal of enabling real-world applications in finance, identity, social media, and AI.

User Engagement Auditing

When payments are tied to volume — content promotion billed by reach, healthcare services reimbursed per patient — clients must trust the platform’s self-reported numbers. VAR introduces a cryptographic primitive that fixes this: the paying party can verify the count of users served without learning individual identities. We demonstrate it with TrueReach, a system for verifiable content promotion on Bluesky.

• Position paper: Verifiable Aggregate Receipts with Applications to User Engagement Auditing (ePrint 2025/2330)
• Demo: TrueReach on Bluesky (https://var.hackingdecentralized.com)

Distributed SNARKs and Proof Markets

Generating ZK proofs for large-scale computations is computationally expensive. Distributing the work across many provers achieves near-linear speedup. We also study pricing and fair allocation of prover work in open markets.

• Cirrus: Performant and Accountable Distributed SNARK (NDSS'26)
• Dinocchio: Distributed Prover for Ring Arithmetic (ePrint 2026/159)
• Prooφ: A ZKP Market Mechanism (FC'25)
• zkBridge: Trustless Cross-chain Bridges Made Practical (CCS'22)

MEV: Measurement and Mitigation

MEV arises when block producers reorder, insert, or censor transactions for profit, harming user fairness and creating centralization incentives. We measure these effects empirically and work on mitigations.

• Geographical Centralization Resilience in Ethereum's Block-Building Paradigms (SIGMETRICS'26)
• Blockspace Under Pressure: An Analysis of Spam MEV on High-Throughput Blockchains (arXiv 2604.00234)
• Boost+: Equitable, Incentive-Compatible Block Building (arXiv 2602.04007)
• Decentralization of Ethereum's Builder Market (S&P'25)
• Insecurity Through Obscurity: Veiled Vulnerabilities in Closed-Source Contracts (arXiv 2504.13398)
• Data Independent Order Policy Enforcement: Limitations and Solutions (CCS'24)

Censorship in Blockchains

Censorship occurs when block producers deliberately exclude transactions — motivated by bribes, regulatory pressure, or competitive advantage. We study mitigations including inclusion lists and multi-proposer BFT protocols.

• Censorship Resistance vs Throughput in Multi-Proposer BFT Protocols (ePrint 2026/126)
• AUCIL: An Inclusion List Design for Rational Parties (ePrint 2025/194)
• He-HTLC: Revisiting Incentives in HTLC (NDSS'23)
• Perils of Parallelism: Transaction Fee Mechanisms under Execution Uncertainty (ePrint 2026/649)

Anonymity in Network and Consensus

Open networks expose participants to traffic analysis and deanonymization. We build anonymity primitives for decentralized systems: ZIPNet, a DC-net protocol for anonymous broadcast (implemented by Flashbots), and Qelect, the first post-quantum SSLE protocol that hides block proposer identities in consensus.

• Qelect: Lattice-based Single Secret Leader Election Made Practical (Security'25)
• ZIPNet: Low-bandwidth anonymous broadcast from (dis)Trusted Execution Environments (PETS'25)

Data Liberation, Provenance and Applications

Town Crier and DECO were among the first to establish verifiable provenance of TLS-encrypted web data — turning any HTTPS site into a source of verifiable claims — initiating the zkTLS area, with protocols now deployed by Chainlink and numerous startups.

• CanDID: Can-Do Decentralized Identity with Legacy Compatibility, Sybil-Resistance, and Accountability (S&P'21)
• DECO: Liberating Web Data Using Decentralized Oracles for TLS (CCS'20)
• Town Crier: An Authenticated Data Feed for Smart Contracts (CCS'16)

Reducing Trust in TEEs

TEEs offer strong confidentiality and integrity but are vulnerable to side channels. We design systems that tolerate, detect, or economically disincentivize leakage.

• ZIPNet: Low-bandwidth anonymous broadcast from (dis)Trusted Execution Environments (PETS'25)
• CrudiTEE: A Stick-and-Carrot Approach to Building Trustworthy Cryptocurrency Wallets with TEEs (AFT'24)
• Sealed-Glass Proofs: Using Transparent Enclaves to Prove and Sell Knowledge (EuroS&P'17)