← Back to home

JudgePass

Eliminate hackathon bias with verifiable AI scoring on blockchain for fair, transparent results.

Problem Statement

JudgePass: Decentralized AI-Powered Judging System The Problem We Solve:** Traditional hackathon and competition judging suffers from critical flaws: human bias, inconsistent evaluation criteria, lack of transparency, time-consuming manual processes, and no way to verify or dispute results. Participants often feel judged unfairly, and organizers struggle with scalable, consistent evaluation across hundreds of submissions.Our Solution:JudgePass is the world's first decentralized, AI-powered judging system that combines blockchain technology, verifiable AI computation, and encrypted smart contracts to deliver completely transparent, bias-free, and cryptographically verifiable judging results.How It Works:Intelligent NFTs (INFTs): Each judge is represented as an NFT containing encrypted metadata that defines specific judging rubrics, evaluation criteria, and scoring methodologies. These INFTs can be owned, transferred, or leased to authorized evaluators.0G Network Integration: We leverage the 0G Compute Network for AI inference, providing access to state-of-the-art models like Llama 3.3 70B and DeepSeek R1 70B running in Trusted Execution Environments (TEE) with cryptographic verification of every response.Encrypted Metadata Storage: Judge configurations are encrypted using AES-GCM encryption and stored on the 0G Storage Network, ensuring tamper-proof preservation of evaluation criteria while maintaining confidentiality.Verifiable AI Scoring: When projects are submitted for evaluation, our system automatically analyzes submissions against predefined rubrics using verifiable AI. Every score includes detailed justifications and cryptographic proofs of computation integrity.Key Features:Bias Elimination: AI judges apply identical criteria to every submission, removing human prejudices and favoritismComplete Transparency: Every scoring decision includes detailed justifications explaining exactly why each score was assignedCryptographic Verification: All AI responses are verified using TEE technology, ensuring results haven't been tampered withScalable Evaluation: Process hundreds of submissions simultaneously with consistent qualityImmutable Results: All scores and justifications are permanently recorded on blockchainFlexible Rubrics: Create custom evaluation criteria for different competition types (technical innovation, user experience, business potential, etc.)Authorization System: NFT owners can authorize specific wallets to use their judges while maintaining ownershipTechnical Architecture:Smart Contracts: ERC-721 based INFTs with encrypted metadata and authorization mechanismsFrontend: React-based dashboard for judge management and submission evaluationBackend: Node.js API with 0G SDK integration for AI inference and storageEncryption: AES-GCM for metadata protection with symmetric key managementBlockchain: Ethereum-compatible networks with 0G Network for compute and storageUse Cases Beyond Hackathons:Grant application evaluationAcademic peer reviewStartup pitch competitionsArt and design contestsJob candidate assessmentResearch proposal reviewsImpact:JudgePass transforms subjective evaluation processes into objective, verifiable systems. By eliminating human bias and providing complete transparency, we create fairer opportunities for participants while reducing evaluation burden on organizers. Every decision is auditable, disputable through cryptographic proof, and consistently applied.Innovation:This is the first system to combine blockchain-based judge ownership, verifiable AI computation, and encrypted metadata management into a comprehensive judging platform. We're not just digitizing traditional judging—we're fundamentally reimagining how evaluation systems can work in a decentralized world.The Future:We envision a world where every important decision—from funding allocation to academic assessment—can be made with transparent, verifiable, and bias-free AI systems while maintaining human oversight and customizable criteria. JudgePass is the foundation for this future of fair and transparent evaluation.

Solution

How JudgePass Was Built?Core TechnologiesSmart Contract FoundationSolidity (v0.8.20)for our INFT (Intelligent NFT) contractOpenZeppelinlibraries for secure ERC-721 implementation, ownership, and reentrancy protectionOur INFT contract extends standard ERC-721 with specialized functionality for encrypted metadata storage and judge authorization, enabling token owners to authorize specific wallets to use their judge NFTs.API ArchitectureBunruntime for TypeScript executionCustom routing systemwith support for path parametersREST APIendpoints for judge metadata, scoring, and scorecard managementWe built a lightweight router that efficiently handles dynamic routes like/judge/:tokenId/metadata, giving us precise control over request handling and response formatting.Frontend ImplementationReactwith TypeScript for type safetyTailwindCSSfor responsive stylingWagmifor wallet connectionsThe frontend provides a complete judging workflow from loading an existing judge NFT to evaluating submissions and displaying AI-generated results.0G Storage Network We integrated with the0G Storage Networkvia the@0glabs/0g-ts-sdkto enable:Encrypted metadata storagewith secure upload and retrievalMerkle tree verificationof data integrityPermanent storageof judging results and scorecards0G Compute Network We leveraged the0G Compute Networkvia the@0glabs/0g-serving-brokerfor:AI service discovery- finding available providers on the networkVerifiable AI inference- generating provably fair evaluation scoresAI model access- utilizing Llama 3.3 70B and DeepSeek R1 70B modelsEncryption SystemFor secure judge metadata storage, we implemented:AES-GCM encryptionfor symmetric key cryptographyWeb Crypto APIfor standards-compliant encryption/decryptionBase64 encodingfor transferable encrypted dataSmart Mock System One of our most innovative implementations is theSmart Mock Systemthat ensures reliability during demos:This system analyzes project descriptions for relevant keywords and generates realistic AI-like responses that provide detailed justifications while clearly marking them as simulations.Architectural Challenges & SolutionsChallenge: BigInt JSON SerializationThe 0G SDK returns BigInt values which can't be directly serialized to JSON. We implemented a transformation layer:Challenge: Judge Mint Workflow We created a specialized mint script that:Encrypts judge metadata with AES-GCMUploads to 0G StorageMints an INFT with encrypted URI and hashCurrent Limitations & Future WorkWhile our prototype demonstrates the core concept, several features are planned for future iterations:Full marketplace implementationfor judge NFT tradingDispute resolution mechanismfor challenging scoresEnhanced frontend visualizationof judging resultsImproved wallet integrationwith multiple providersThe most challenging aspect of the build was integrating with the 0G network while handling connectivity issues and ensuring our demo could proceed reliably with our smart mock system.This combination of technologies enabled us to create a working prototype of a decentralized judging system with verifiable AI that solves real problems in hackathon evaluation.

Hackathon

ETHGlobal New Delhi

2025

Contributors