Automated Quantitative Portfolio & Risk Engine (quant-matrix-allocator)

An end-to-end, serverless quantitative asset allocation and risk management engine. It automatically optimizes multi-asset portfolio weights using Mean-Variance Sharpe Maximization, computes historical Value at Risk (VaR 95%), runs automated unit test suites (pytest), and renders updated visual allocation artifacts daily via GitHub Actions.

---

📐 Mathematical Formulation

1. Portfolio Return & Volatility

• Expected Return: $E[R_p] = w^T \mu$
• Portfolio Variance: $\sigma_p^2 = w^T \Sigma w$

2. Sharpe Ratio Optimization

$$\max_w \frac{w^T \mu - R_f}{\sqrt{w^T \Sigma w}} \quad \text{s.t.} \quad \sum_{i=1}^{N} w_i = 1, \quad 0 \le w_i \le 1$$

3. Historical Value at Risk (VaR 95%)

$$VaR_{0.95} = -\text{Percentile}_{5}(R_p)$$

---

📊 Live Visual Artifact

How to Read This Artifact

• Optimal Weights: Bars reflect portfolio allocation calculated via SciPy's Mean-Variance Optimization (scipy.optimize.minimize) aiming to maximize the Sharpe ratio ($0.83$) under long-only constraints ($\sum w_i = 1, w_i \ge 0$).
• Risk Profiles: Zero-weight allocations (Global Bonds, Crypto Index) indicate sub-optimal risk-adjusted yield relative to the covariance structure during optimization.
• Data Source & Pipeline: Inputs are generated from a multivariate daily returns matrix ($N=1000$ simulation steps) calibrated against historical asset covariance matrices. $95%$ Historical VaR ($1.25%$) is computed directly from the 5th percentile of simulated portfolio return distributions ($VaR_{0.95} = -\text{Percentile}_5(R_p)$).

---

🛠️ Tech Stack & Quality Gates

• Core Analytics: Python 3.10, NumPy, Pandas, SciPy
• Visualization: Matplotlib, Seaborn
• Testing & CI/CD: Pytest, GitHub Actions