Documenting the training of a RL agent to provide price bounds for concentrated liquidity providing on UniV3 using recent price action. The agent is trained to predict the smallest range with the maximum estimated lifespan.
Backtesting on real-world scenarios with Impermanent Loss and Fee accrual is necessary before using this for any real world use case.