ZK-XGen-AI — Privacy Infrastructure Platform

🖥️

Best Viewed on Desktop

The technical documentation and app previews are optimized for desktop viewing.

📄

Academic Papers

ZK-Sentinel originated from doctoral research at Universidad Nacional de La Plata. Three peer-reviewed papers spanning privacy, security, and exploit verification.

Core Paper — V5

ZK-Sentinel: Coercion-Resistant Financial Privacy with Selective Compliance through Dual-Logic Zero-Knowledge Circuits and On-Chain Machine Learning

ZK-XGen-AI Research · Universidad Nacional de La Plata

32 pages, 12 tables, 19 references. Presents the complete protocol with formal definitions, security proofs (Theorems 1-4), experimental evaluation on Sepolia, and comprehensive comparison with existing solutions.

Six Contributions (Paper §1.4)

C1Dual-Logic ZK Circuit Architecture — first implementation of mathematically indistinguishable real/decoy selection in a zk-SNARK

C2ZKML for Coercion Detection — first integration of ML coercion detection within a zero-knowledge proof system via EZKL-compiled Halo2

C3Privacy Pools with Selective Compliance — PoI via ASPs, partial withdrawals with private change (13,624 constraints), selective disclosure (22,457 constraints)

C4Partial Withdrawals with Private Change — change cryptographically indistinguishable from regular deposit, preserving anonymity set

C5Nine-Layer Defense Model — 4 ZK circuits, 5 verifiers, 92 pools, 6 privacy infrastructure invariants, gasless withdrawals, fee model (89/8/3)

C6Production Implementation — deployed on Sepolia with E2E validated TXs for all operations. Zentinel-Audit (28 tools): 63 alerts → 1 low-severity finding

Under Review32 pages19 References

Doctoral Thesis

Applications of Generative AI in Distributed Architectures for Blockchain Security and Zero-Knowledge Cryptography

Universidad Nacional de La Plata · Defense: 2026-2027

Two main contributions: (C1) Zentinel-Audit — smart contract security analysis with 28 tools, GAEV exploit generation, and multi-LLM agents. (C2) ZK-Sentinel — privacy protocol with 10 layers including the novel Shadow Passphrase, dual nullifier architecture, and ZKML coercion detection.

C2: Zentinel-AuditC3: ZK-Sentinel

Grant Applications

EF Fellowship & Arbitrum Grants

Shadow Passphrase and the "peace of mind = cryptographic parameter" concept positioned as unique differentiator. The anti-coercion mechanism addresses a real-world threat (189+ documented attacks, $84M+ losses) not covered by any existing privacy protocol.

Key References

Academic Foundation

[6]  Pertsev, Semenov, Storm — Tornado Cash Privacy Solution (2019)
[8]  Hopwood, Bowe, Hornby, Wilcox — Zcash Protocol Specification (2016)
[10] Buterin, Illum, Nadler, Schär, Soleimani — Privacy Pools (2023)
[11] 0xbow — Privacy Pools: Compliant Privacy on Ethereum (2024)
[12] Goldwasser, Micali, Rackoff — Knowledge Complexity (1989)
[14] Groth — On the Size of Pairing-Based Arguments, EUROCRYPT (2016)
[15] Grassi et al. — Poseidon Hash, USENIX Security (2021)
[17] EZKL Team — Easy Zero-Knowledge Machine Learning (2023)
[18] Juels, Catalano, Jakobsson — Coercion-Resistant Elections (2005)