I created a single RL agent(stocks trading agent) for a leading financial institution, GTCO trading in NGX(Nigeria Exchange Group), and eval the agent ROI on a 30day's performance
I sourced for historical stock price via https://www.investing.com/equities/guaranty-bnk-historical-data
And used the use gym standards for vanilla RL architecture
Set up the Q_learning env
Trained and eval the agent across 500 episodes
And checked agent ROI performance for a 30-day window