ZKward
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

ZKward is built on a simple premise: AI shouldn't make up numbers, especially when real money is involved. While most AI finance tools quietly ask you to trust unverified model outputs, ZKward pairs Meta's open-weight Muse Spark Open with zero-knowledge attestations to ensure every financial claim is cryptographically proven and tied directly to live data across 13 languages. By running a live USDC community pool on Sui Mainnet, test pools on Solana, and simulated trading books that rigorously account for every transaction fee and slippage cost, we effectively democratize hedge-fund-grade risk management for everyone. By fusing real-time odds from prediction markets directly into our signal engine, our autonomous agent actively manages asset risk—answering complex queries about real-world performance, net asset value, and signal direction with complete, uncompromising transparency.