DarwinFi
DarwinFi's Team
Problem Statement
Autonomous agents managing real money need transparent, auditable decision-making with built-in risk controls. Current trading bots use static strategies that degrade over time. DarwinFi solves this by creating a self-improving system where 12 strategies compete in a Darwinian tournament. Performance is scored by a weighted composite (PnL, Sharpe ratio, win rate, drawdown). Strategies that outperform get promoted to live trading; underperformers get evolved or replaced. Per-strategy budget isolation in DarwinVault smart contracts caps downside risk. The entire evolution process is logged as a conversation audit trail.
An autonomous, self-evolving crypto trading agent that uses Darwinian competition between 12 concurrent strategies on Uniswap V3 (Base). The top-performing strategy trades live on-chain; the rest paper trade and compete to dethrone it. Strategies evolve via AI (Claude + Venice AI) analyzing performance metrics and generating parameter variations across three roles: Mutant (creative exploration), Tuner (conservative optimization), and Hybrid (trait synthesis). Three main bots -- Apex (momentum), Viper (mean-reversion), and Blitz (breakout) -- each with three variations, creating a population of 12 that continuously improves through natural selection.
Build Timeline
Team
DarwinFi
admin
MandateIncrease your chances to win
- ›Most agents in the hackathon are exposed to prompt injection
- ›This might cause overspending and loss of funds
- ›Security is a crucial part of the hackathon
Share on X
Tell the world about this project
Tracks
Intention
Plans to continue
DarwinFi is designed as a long-running autonomous trading system. After the hackathon, the plan is to deploy contracts on Base mainnet, fund the vault with real USDC, and let the Darwinian competition run continuously. The self-evolving architecture means the system improves over time without manual intervention.