MSc candidate in Economics and Finance (Quantitative Finance), University of Milano-Bicocca
Expected graduation: March 2027 · Based near Milan, Italy
I'm interested in portfolio construction, systematic asset allocation and investment research, with a second focus on derivatives and quantitative market risk. This page collects completed university research projects, with their source code, figures and limitations. They are academic studies, not live investment strategies.
1. Adaptive Asset Allocation · MATLAB · Group project
Compared 21 multi-asset specifications across 10 market indices: inverse volatility, risk parity, maximum diversification, momentum and volatility targeting. The weekly adaptive strategy includes modelled transaction costs. The repository documents the in-sample design, the lack of an independent holdout and implementation limitations.
2. Optimal Capital Allocation · MATLAB · Group project
Applied VaR and Tail Conditional Expectation to capital allocation for a five-equity portfolio, including alternative tail-state weighting schemes and an 800-observation rolling attribution exercise. The rolling exercise is retrospective rather than a leakage-free out-of-sample backtest.
3. Financial Time-Series Econometrics · R · Individual coursework
Twenty exercises covering ARMA/ARIMA, ARCH/GARCH/EGARCH, volatility forecasting, VaR and Expected Shortfall. The repository includes code, original figures and an audit of modelling and implementation details.
4. Advanced Derivatives Market Analysis · MATLAB · Group project
EURO STOXX 50 options, no-arbitrage checks, implied volatility, Black-Scholes versus Monte Carlo pricing, EUR OIS curve construction and a callable credit-equity-linked bond. The README separates classroom approximations from production methodologies.
5. Financial Market Risk Modelling · R · Group project
Newey-West and AR-GARCH inference, Variance-Gamma marginals with a Gaussian copula, and a synthetic credit-loss simulation using Student-t dependence. The credit exercise is methodological, not calibrated to a real loan portfolio.
Each project links to its own code and explains the data requirements. Some raw market datasets are not publicly redistributed because they were provided under third-party or academic access conditions. Where applicable, AUDIT.md records limitations or corrections rather than hiding them.
The work above was completed either individually or in a university group, as labelled. The original collaborators are acknowledged in the individual repositories. Historical backtests and synthetic simulations are not live performance.
Tools used: MATLAB, R, Git/GitHub; Python (NumPy, pandas) for data work. Academic exposure to Bloomberg Terminal and FactSet.
I'm also working on a master's thesis research design concerning 21-day SPY tail-risk forecasting. It is a research plan, not a completed empirical project; it is therefore not included among the five completed repositories above.