Open to opportunities

Tommaso Guerrini

@tommasoguerrini

Message

Data science leader specializing in tabular ML, model productionalization.

Italy

Message

What I'm looking for

I seek senior roles leading ML initiatives where I can build production-ready tabular models, mentor teams, and shape ML lifecycle and CI/CD in collaborative, engineering-driven environments.

I am a data science leader with nine years of experience building and productionizing tabular machine learning solutions across finance, insurance, automotive, manufacturing and telco. I have led teams, acted as product owner for model releases, and implemented end-to-end ML pipelines using tools like Kubeflow, MLflow, KServe and ArgoCD while mentoring junior colleagues.

My work includes developing ensemble PD models, anomaly-detection frameworks that reduced onboarding time, graph-based fraud features, and feature stores orchestrated with Airflow and Jenkins; I am proficient in Python, PySpark and experienced with GCP and AWS, and I actively contribute to Kaggle and other data-science communities.

Experience

Work history, roles, and key accomplishments

Current

Tabular Data Lead

Current

CardoAI

Jul 2022 - Present (3 years 5 months)

Led tabular data modeling and productionization for credit risk and data quality, developing ensemble PD models for US consumer and Italian SME markets and implementing automated retraining and monitoring pipelines to improve reliability.

LightGBM XGBoost SHAP Kubeflow MLFlow Airflow KServe Pyspark Python

Machine Learning Scientist

Generali Italia

Jan 2020 - Jun 2022 (2 years 5 months)

Developed and deployed fraud detection and text classification models on Google Cloud, improving fraud model accuracy ~15% via graph-based features and delivering reusable feature pipelines and CI/CD standards.

Google Cloud Platform BigQuery Pyspark LightGBM Optuna Neo4j Airflow Jenkins Python

Data Scientist

Data Reply

Aug 2016 - Dec 2019 (3 years 4 months)

Built forecasting and anomaly-detection models across automotive, manufacturing, telco and media clients, achieving >10% improvement in TV audience forecasting and delivering supply-chain and demand-planning solutions.

Pyspark XGBoost LightGBM Time Series Forecasting Docker AWS Google Cloud Python

Education

Degrees, certifications, and relevant coursework

Politecnico di Milano

Master of Science, Applied Statistics

2014 - 2017

Grade: 110/110

Completed a Master of Science in Mathematical Engineering with a focus on Applied Statistics; final project on machine learning models for forecasting daily supermarket sales.