Особисті дані приховані

Daniil Hutsol
CONTACT
Phone: [відкрити контакти](див. вище в блоці «контактна інформація»)
Quantitative Methods Student, SGH Warsaw School of Economics Email: [відкрити контакти](див. вище в блоці «контактна інформація»)
Location: Warsaw, Poland

SUMMARY
Quantitative Methods and Information Systems student who has spent the past year turning real-world data into working models,
from predicting corporate bankruptcy to optimizing investment portfolios.
RELEVANT EXPERIENCE & PROJECTS
Math Tutor | Private students & peer exam support, 2026
Tutored high school students in mathematics; supported university classmates across all higher mathematics courses.
Portfolio Optimizer | Python, SciPy, yfinance
150+ ETF universe, 2020–2026, train/test split
Built a portfolio optimization pipeline combining Monte Carlo simulation with numerical optimization for a barbell asset allocation
strategy:
• Simulated 300,000 candidate portfolios via Monte Carlo, then refined with SciPy SLSQP to maximize Sharpe ratio under barbell
allocation constraints (65–75% conservative, 25–35% high-yield), fitting weights on a training period and validating on a held-out
test period
• Achieved an out-of-sample Sharpe Ratio of 1.09 (17.5% annualized return, 11.9% volatility, -10.1% max drawdown) on the test
set, a 9.7% degradation from the 1.21 in-sample Sharpe, indicating limited overfitting
• Evaluated downside risk via Sortino, Calmar, and VaR/CVaR across an automated, 150+ ETF universe pulled and filtered via
yfinance
Bankruptcy Prediction System | Python, XGBoost, Optuna, SHAP
Polish Companies Dataset (2000–2013)
GitHub: github.com/lzanam1/bancrucy
Built an ML pipeline to predict corporate bankruptcy for bank credit risk screening:
• Achieved ROC-AUC 0.86 and 90.9% recall on an imbalanced dataset (~5% bankruptcy rate)
• Projected a 57.5% reduction in analyst workload on held-out test data by automatically clearing low-risk companies
• Implemented MICE + RandomForest imputation for features with up to 39% missing values
• Tuned XGBoost hyperparameters via Optuna (50 trials, F2-score objective); outperformed a Random Forest baseline on ROC-
AUC, recall, and PR-AUC at a matched decision threshold
• Applied SHAP analysis confirming economic consistency of model decisions (profitability, liquidity, debt coverage)
EDUCATION
SGH Warsaw School of Economics
Bachelor in Quantitative Methods in Economics and Information Systems
Expected Graduation: 2027
Major GPA: 4.68/5.0
Relevant coursework:
Probability Theory, Econometrics, Induced Decision Rules (Machine Learning), Microeconomics,
Computer Programming, Accounting, Algebra, Optimization Methods, Statistics
Additional:
Mathematical Economics (Ordinary Differential Equations); independent study of stochastic calculus
TECHNICAL SKILLS
Python: pandas, NumPy, scikit-learn, XGBoost, Optuna, SHAP, Matplotlib, SciPy, yfinance
SQL: joins, aggregation, nested subqueries
R: tidyverse (dplyr, ggplot2), decision trees, random forests
Microsoft Excel: data analysis, spreadsheet-based simulations
Power BI: data visualization, dashboards
LANGUAGES
English (C1)
Polish (B1)
Ukrainian (Native)
German (A1/A2 – Basic)
Russian (Native)