An Enterprise-grade Risk Management Platform designed for multi-asset portfolio modeling.
- HPC Monte Carlo Engine (C++17): Developed a custom simulation module integrated via Pybind11. Utilizing OpenMP for full CPU core parallelization and SIMD Optimization, achieving 400x+ speedup over standard Python loops for 50,000+ path Geometric Brownian Motion (GBM) simulations.
- Hybrid Volatility Forecasting: Implemented a unique pipeline where GARCH(1,1) econometric models are corrected by LSTM Neural Networks (PyTorch) to capture non-linear market anomalies and volatility clustering.
- Dynamic Correlation (EWMA): Real-time covariance matrix updates to detect "correlation breakdown" during market stress events.
- Statistical Backtesting: Automated validation layer using Kupiec (POF) and Christoffersen independence tests to ensure VaR model integrity.
- Optimized Data Layer: Custom SQL caching layer and memory-aligned data structures to minimize cache misses and reduce data retrieval latency by 95%.
- Languages:
C++17/20(Templates, Metaprogramming),Python 3.x(AsyncIO). - Optimization:
OpenMP(Multi-threading),SIMD(Vectorization),Pybind11(Zero-copy memory transfer). - Backend & Dev:
FastAPI,Docker & Compose,PostgreSQL/SQLite.
- Libraries:
PyTorch(LSTMs),NumPy,Pandas,SciPy,Plotly. - Focus Areas: Tail Risk Estimation (Parametric, Historical, Monte Carlo VaR/ES), Portfolio Optimization (Mean-Variance), Spectral Dynamics.
| GitHub Stats | Top Languages |
|---|---|
- Developing Real-time pricing engines for derivative instruments.
- Researching Spectral Dynamics in high-frequency financial time series.
- Refining Institutional Dynamic Weight Bounds in portfolio rebalancing.