AI Trading Strategies Research
Five quantitative research projects completed for the AI Trading Strategies Nanodegree, covering risk modeling, feature engineering, portfolio construction and machine-learning trading agents.
Architecture
Geometric Brownian motion calibration with expected shortfall on index data stored in SQLite; a feature pipeline over prices and macro series; a risk-parity futures portfolio with drawdown analysis; a deep Q-network trading agent on Bollinger Band features; and a random forest that predicts five-day direction from price, volatility index and search-trend data.
Highlights
- Verified Nanodegree, August 2026
- Deep Q-network trading agent
- Risk parity and expected shortfall