Senior Data Scientist

Development of mathematical models for electricity production and load. Development and optimization of trading algorithms in Python (numpy/Pandas/scikit-learn): increased trading profitability by 100 basis points in first 60 days. Authored comprehensive reports in Tableau, maintained and improved data pipelines from MongoDB and PostgreSQL.

Played a key role in shaping regional government strategy through advanced statistical analysis for strategic planning. Led the construction of data pipelines and authored comprehensive reports in Oracle BusinessObjects, showcasing expertise in data visualization and reporting. Engineered algorithms in R for efficient sampling, data analysis, and model building, demonstrating a data-driven approach to decision-making. Published groundbreaking research on policy evaluation modeling using advanced statistical methods (Springer Genus, 2023). Developed demographic forecasting models and implemented innovative data cleaning methods for improved data quality. Designed and implemented robust sampling schemes, supporting data collection, analysis, and internal audit procedures.

Created a team ranking system for the online football manager game Hattrick.org. Developed a customized rating system based on ELO ratings, tailored to the unique context of the game. Performed calculation of approximate expected values of nested stochastic processes, integrating them into the game code. Utilized various data analysis methodologies, including simulation, generalized linear models, regression trees, and custom dissimilarity measures, in Python and R.

Instructed ~40 students per year in descriptive statistics for Statistical and actuarial science students (2008-2019). Instructed 150+ students per year in descriptive statistics, probability calculus and inferential statistics for Economics students (2016-2021). Development of R – Shiny applications to facilitate statistical learning for high school and junior students.

Developed hierarchical Bayesian models in Python, R and C based on matrix factorization (Latent Dirichlet Analysis, an upgrade of PCA) to classify unstructured textured data by topic, cluster analysis (KMeans) by topic and automatic classification by topic. Received the 2007 InnovAction prize for the most innovative companies in the Friuli Venezia Giulia region for the project of a smart search engine on administrative procedures data on which the Master thesis was based.