Simone De Bonis
Data Scientist · Torino, Italy
I design and deploy production machine learning systems and quantitative forecasting models. Currently working as Data Scientist at Intesa Sanpaolo in Torino, after building energy trading forecasting pipelines at Psaier Energies.
Mid-term empirical assessment of United Nations Sustainable Development Goals (Goals 8, 9, and 11) through interactive Tableau dashboards built on Our World in Data.
Understanding the distinction between heavy-tailed, fat-tailed, and light-tailed probability distributions and their implications for the Central Limit Theorem.
What I’m Working On
Full CV →Data Scientist
Engineering, scaling, and industrializing machine learning models and data-driven pipelines within Italy’s leading banking group.
Data Scientist (Energy Trading)
Worked in the energy trading department developing forecasting models and cloud data pipelines for the Italian energy market:
- Developed and monitored statistical and ML models for Italian energy market forecasts and analysis.
- Migrated legacy data workflows from on-premises infrastructure to Microsoft Azure.
- Analyzed large-scale market and consumption datasets using Python and SQL.
- Built real-time Power BI dashboards to support energy trading decisions.
Recent Work & Research Highlights
All projects →Master’s Thesis — Automotive Predictive Maintenance
Developed a Python ensemble framework integrating knowledge from multiple machine learning models on telemetry data to predict and explain engine failures for a leading automotive manufacturer.
Multivariate Time Series Anomaly Detection
Applied advanced machine learning techniques for multivariate anomaly detection on high-frequency time series extracted and preprocessed from enterprise server logs.
Early Warning System for Italian Municipalities
Collaborated on an early warning classification system using Support Vector Machines (SVM), Multi-Layer Perceptrons (MLP), and class-balancing techniques (SMOTE & ROSE) to predict financial default in Italian municipalities.
Latest Articles
View Full Archive →Fat and Heavy Tails: Differences and Implications
Understanding the distinction between heavy-tailed, fat-tailed, and light-tailed probability distributions and their implications for the Central Limit Theorem.
Outliers, Extremes and Black Swans
Distinguishing measurement anomalies from genuine extreme realizations and structural tail events in quantitative modeling.
Intuition and History of Matrix Multiplication
From standard cubic complexity O(n³) and Strassen's 1969 breakthrough to AlphaTensor and reinforcement-learning discovered algorithms.
One Key Difference Between Machine Learning and Statistics
Comparing the unbiasedness constraint (BLUE / MVUE) in classical econometrics with the bias-variance trade-off in supervised machine learning.
K-Nearest Neighbors (KNN): Probabilistic & Geometric Intuition
Deriving Voting KNN and Volumetric KNN from non-parametric density estimation and Bayes classification, plus practical efficiency improvements.
