Build a poker advice tool using MCCFR

Job ID: 33767575

Budget: €750 – €1,500 EUR

I am looking for someone that can code a poker advice tool for poker cash games and tournaments that uses the Monte-Carlo Counterfactual Regret Minimisation model. The goal is to have a strategy reaching equilibrium and thus not being exploitable.

The following steps should be taken:
1. Create the poker environment where the AI can play multiple agents against itself.
2. Give each agent different opening hand ranges and strategies to use and learn from.
3. Train the model using a standard Texas No Limit Hold'em cash game
4. Train the model using multiple tournament environments.
5. Use Neural Networks to identify online poker tables to read from using image recognition
6. Create a program that can read the poker table and give advice on what action to take based on strategy and situation
7. Let the model learn from the advice given if follow.

Knowledge about No Limit Hold'em and ranges can help a lot for developing this project.
I found a couple articles and code regarding this topic and i think these can help as well.

The final product should be a program with an easy to use GUI, an optimal poker strategy for both cash games as tournaments, and the program should be able to read poker tables from a different device using teamviewer or something similar.