Back to Projects
AI Agent

Poker AI Agent (CFR)

A comprehensive poker game application featuring AI opponents powered by Counterfactual Regret Minimization (CFR) algorithms. The system features real-time multiplayer functionality, opponent modeling, and risk management strategies.

Source Code

Core Technology

Node.jsTypeScriptCFR AlgorithmSocket.ioMongoDBReactVite

System Design

Strategic AI Decision Engine

  • •Implemented Counterfactual Regret Minimization (CFR) for approximating Nash Equilibrium in imperfect information games.
  • •Real-time Opponent Modeling system that adapts to player aggression and bluffing frequencies.
  • •Event-driven socket architecture to synchronize game state between human players and AI agents.

Architecture

Hybrid architecture combining a Node.js game server with a dedicated AI decision engine. The AI engine runs CFR algorithms in real-time to compute optimal strategies based on game state and opponent history.

Architecture Diagram Placeholder

Key Features

01

CFR-based decision making

02

Opponent Modeling & Adaptability

03

Bluffing detection & execution

04

Real-time game state synchronization