Macro-Systemic Alpha Signals Engine

By Dawit Yimer Gebreegziabehair, MScFE | WorldQuant BRAIN Gold Medalist

What it is

This is a fully automated, multi-strategy pipeline built for the Numerai Signals tournament. It analyzes liquid US equities (S&P universe) to extract statistical arbitrage opportunities, while simultaneously monitoring the U.S. Federal Reserve for macroeconomic regime shifts.

How it works

• Macro Shield (Systemic Risk): The engine maintains an authenticated API connection to the St. Louis Fed (FRED) to monitor the 10Y-2Y Treasury yield spread. If the yield curve inverts (indicating recessionary risk), the system automatically scales down signal conviction by 50% to preserve capital.
• Alpha Logic (Statistical Arbitrage): It executes a volatility-adjusted mean-reversion strategy. The model calculates 20-day Z-scores across the equity universe, penalizing highly volatile assets to ensure stable, risk-adjusted returns.
• Market Neutrality: Final signals are ranked and normalized to remain uncorrelated with standard market beta, prioritizing Feature Neutral Correlation (FNCv4) and Meta Portfolio Contribution (MPC).
• Infrastructure: The entire pipeline is 100% serverless. Hosted on GitHub Actions, it handles ticker mapping, data synchronization, inference, and API delivery to the hedge fund every 24 hours without human intervention.

Tools Used

• Python 3.11 (Core Architecture)
• FRED API (Macroeconomic Telemetry)
• Yahoo Finance (Equity Price-Action)
• Polars & NumPy (High-speed Data Orchestration)
• GitHub Actions (Daily MLOps Execution)

Live Performance

Daily signals, Meta Portfolio Contribution (MPC), and live reputation scores are tracked directly by the hedge fund on the Numerai platform:
DAWITYIMER (Signals Dashboard)

---

MIT License | Institutional Quantitative Research