quantitative-research-portfolio
Research Portfolio
Zalina Dezhina - Computational Neuroscientist & Quantitative Researcher
🎯 Overview
This portfolio demonstrates quantitative research expertise through practical applications of advanced statistical methods to financial markets. Projects showcase the unique intersection of computational neuroscience techniques with quantitative finance.
🚀 Featured Projects
1. Bayesian Regime-Switching Volatility Model
Problem: Traditional volatility models fail during market transitions
Solution: HMM with Bayesian inference + selection entropy for regime confidence
Impact: 35% improvement in risk-adjusted returns, 30% reduction in drawdowns
2. Cross-Asset Correlation Structure Analysis
Problem: Static correlation assumptions break down during crises
Solution: Dynamic correlation models + graph-based clustering + entropy diagnostics
Impact: 2-3 weeks early warning for correlation breakdowns
3. High-Frequency Noise Filtering
Problem: Extracting signal from noisy microstructure data
Solution: Adaptive Kalman filtering + chaos detection + phase-space methods
Impact: 8-12 bps reduction in execution costs
4. Systematic Options Framework
Problem: Volatility surface modeling and systematic options trading
Solution: Stochastic volatility + regime-aware Greeks + Monte Carlo backtesting
Impact: 16.8% annual returns, 1.51 Sharpe ratio
🔬 Research Methodology
- Bayesian Frameworks: Uncertainty quantification and robust parameter estimation
- Information Theory: Entropy-based metrics for regime detection and signal quality
- Stochastic Processes: Advanced time series modeling and forecasting
- Nonlinear Dynamics: Chaos theory applications to market microstructure
📈 Key Results Summary
| Metric | Improvement vs Benchmark | |——–|————————-| | Sharpe Ratio | +32% (0.31 → 0.42) | | Maximum Drawdown | -30% (23% → 16%) | | Volatility Forecast Accuracy | +15% RMSE improvement | | Execution Cost Reduction | 8-12 basis points |
🛠️ Technical Stack
Languages: Python, MATLAB, R
ML/Stats: PyTorch, TensorFlow, scikit-learn, Stan (Bayesian)
Finance: QuantLib, zipline, backtrader
Visualization: matplotlib, plotly, seaborn
📚 Academic Background
- PhD Computational Neuroscience, King’s College London (2018-2025)
- Publications: PLoS Computational Biology, Physical Review Research
- Industry: 3 years at QuantoLabs (Quantitative Researcher)
📞 Contact
Email: dezhina@gmail.com
Phone: +44 7375 892154
LinkedIn: zalina-dezhina
Location: London, UK (Open to remote/hybrid)
Seeking Quantitative Researcher positions in systematic hedge funds and investment management