I gave Claude Code $10,000 and one job: build a quantitative trading strategy from scratch and execute it live. In this video, I walk you through the full build—from the stock screener to the execution engine using the Alpaca API.

Most people get “overfitted” results with AI bots (that fake 80% win rate). I’ll show you the exact prompt I used for walk-forward optimization that turned a “too good to be true” bot into a realistic, profitable system.

What we cover:

The Prompt: The exact logic I gave Claude to analyze 200+ S&P 500 stocks.

The Strategy: VWAP crossovers, EMA momentum, and volume filters.

The Reality Check: Why Wednesday was a red day and how the bot handled it.

The Tech Stack: Claude Code, Python, SQLite, and Alpaca.

⚠️ Disclaimer: This is for educational purposes. Trading involves risk. I am a Computer Science student and Quant enthusiast, not a financial advisor. Always paper trade before using real capital.

In this video, I walk you through the entire process of building and testing a stock trading AI. Learn how I engineered a daily stock screener that connects to market data APIs, turning $10,000 into real trading results. Discover how this automated trading bot scanned over 200 stocks, built its own strategy, and backtested it against a full year of market data.