Mini Wall Street — Multi-Agent Trading Research System

Mini Wall Street is an algorithmic trading research system that I am building as founder and lead engineer. It uses nine specialized agents coordinated with LangGraph, local Ollama models, and the Codex API to research and evaluate trading decisions. I design the agent workflow, connect market data and execution interfaces for crypto, forex, and US equities, and work on evaluation and risk controls. The system is in demo and paper testing. It has not been deployed with real capital or presented as a profitable strategy. My contribution: system architecture, Python implementation, multi-agent orchestration, API integrations, testing, and product direction. Stack: Python, LangGraph, Ollama, Codex API, CCXT/Binance, MT5, Alpaca.