The problem SwarmSense: Agentic Grant Evaluation Protocol solves

A fully decentralized grant proposal evaluation system that combines Coinbase Developer Platform (CDP) for blockchain interactions with GAIAnet Swarm's decentralized AI models. The system leverages multiple decentralized components to ensure fair, transparent, and automated evaluation of blockchain project proposals with seamless on-chain execution capabilities.

The Problem It Solves

• Decentralized Grant Evaluation: Traditional grant evaluation systems are centralized and often biased. Our solution uses a combination of decentralized AI models and popular SOTA models (GAIAnet Swarm and XAI) to provide fair and transparent evaluations.
• Automated Milestone Tracking: Smart contract-based milestone evaluation system that automatically triggers token distributions and rewards based on AI-powered assessments.
• Multi-Model Consensus: Utilizes both GAIAnet Swarm and XAI models to achieve consensus in proposal evaluation, ensuring more reliable and unbiased results.
• Seamless Blockchain Integration: Leverages Coinbase Developer Platform (CDP) for simplified on-chain actions like token creation, ETH transfers, and NFT minting, making the system more accessible and efficient. We also use Agentkit to allow the orchestration of function/tool calling.

Challenges I ran into

AI Model Integration:

• Implementing integration between GAIAnet decentralised Swarm models such as polygon.us.gaianet.network and llama8b.us.gaianet.network and XAI models
• Created a robust evaluation system that combines scores from multiple AI models and averages them with low latency is hard as, decentralised models are often high latency outputs and can take time
• Developed standardized scoring criteria across different models

Automated Execution:

• Building reliable automated execution of on-chain actions using CDP SDK
• Implementing secure threshold mechanisms for automated token creation and transfers

Additional Features

Added automated token creation using CDP's token deployment features, Created milestone-based evaluation and reward distribution system, and Implemented dual AI model evaluation system using GAIAnet Swarm and XAI