class DataScientist:
def __init__(self):
self.name = "Fabrizio Di Sciorio, PhD"
self.role = "Senior AI Data Scientist @ FiberCop"
self.location = "Rome, Italy 🇮🇹"
self.focus = ["AI & ML", "Real Estate AVM", "ML Forecast", "Overfitting"]
self.currently_exploring = "Long Memory Processes & Fractal Finance"
self.off_duty = ["Cycling (eSRT Team - indoor) 🖥️", "Cycling (Di Sciorio Cycling Team - road) 🛣️"]
def say_hi(self):
print("Thanks for stopping by — let's build something data-driven!")- 🔭 Refining AVM ensemble architectures for real estate valuation
- 📊 Overfitting diagnostics for gradient boosting models (LambdaGuard)
- ⏱️ Multi-horizon time series forecasting techniques
- 📖 Long Memory Processes & Fractal Finance research
When I'm not training models, I'm training legs and lungs.
🚴 Cycling — road & e-sports
- 🖥️ Indoor / e-Racing — riding for eSRT Team
- 🛣️ Road — riding for Di Sciorio Cycling Team
🧠 Same mindset as data science: pace yourself, trust the data (power, heart rate, splits), avoid overfitting your training plan to one good race
|
🤖 AI & Machine Learning
|
🏠 Real Estate AVM & Predictive Analytics
|
📈 ML Forecasting & Diagnostics
|
Core Libraries: Pandas NumPy Polars LightGBM XGBoost CatBoost Optuna Statsmodels Plotly Streamlit
Proficiency
Python ██████████████████░░ 80%
Gradient Boosting (LGBM/XGB/CatBoost) ██████████████████░░ 90%
Time Series Forecasting █████████████████░░░ 85%
R ██████████████░░░░░░ 90%
SQL ████████████████░░░░ 50%
2025 ─● FiberCop — Senior AI Data Scientist
AI Forecasting Systems · EBITDA Analytics · Credit Risk Assessment · Monte Carlo
2020 ─● Prelios — Data Scientist
Ensemble AVM · Real Estate Analytics · Collateral Monitoring
2019 ─● DEMOCOM — Junior Data Scientist
Statistical Analysis · Predictive Analytics · Reporting Automation
2017 ─● Enel Group — Quantitative Analyst Intern
Volatility Modeling · Financial Time Series Analysis
|
🛡️ LambdaGuard Detects the exact moment when boosting algorithms stop learning signal and start memorizing noise.
|
🏠 Real Estate AVM Engine Ensemble stacking model for residential property valuation (80–85% accuracy), with dynamic collateral revaluation at micro-territorial level.
|
|
🎲 CAPEX Monte Carlo Risk Engine Probabilistic framework for investment risk: schedule, cost, sensitivity & scenario analysis.
|
📉 Portfolio Simulation Engine Monte Carlo engine for portfolio analysis and investment simulations.
|
|
🌊 Physical Risk Damage Estimation Copula-based methodology estimating economic damages from floods, landslides, seismic events & storms.
|
|
Financial Market Efficiency Long Memory Processes Fractional Brownian Motion Fractal Finance Hurst Exponent Time Series Forecasting Complex Systems
PhD in Quantitative Finance & Econometrics — Universidad de Almería, Cum Laude
"Estimating Information Inefficiency in Financial Markets Under a Fractional Regime"
MIT Professional Education — Applied Data Science Program
Final Project: Facial Emotion Detection using CNNs and Vision Transformers
- 📄 A Composite Index for Measuring Stock Market Inefficiency — Complexity, 2022
- 📄 Option Pricing under Multifractional Process and Long-Range Dependence — Fluctuation and Noise Letters, 2021
- 📄 Forecasting VIX with Hurst Exponent — Methods and Applications in Fluorescence, 2022
📖 Full publication list on Google Scholar →
