Services.
Some of my previous works.
As a data scientist at The Goodyear Tire & Rubber Company, I developed innovative solutions based on machine learning and optimization, leveraging large and multi-fidelity data sources to support the virtual tire development process. Furthermore, I worked on Tire intelligence projects, designing and developing algorithms to analyze large amounts of data captured from tire sensors to provide real-time proactive solutions for truck fleets.
-
Responsible for the planning, coordination, execution, deployment, and monitoring of data-driven engineering projects involving a cross-functional global team from the USA and Europe.
-
Developing ML models to support advanced tire simulation capabilities leveraging multi-fidelity data
-
Creation of novel algorithms combining multi-performance optimization, FEA simulations, and ML to support the virtual tire design process.
-
Developing ETL pipelines to process semi-structured data by ingesting millions of records from multiple sources to the cloud data lake storage within a CI/CD process.
-
Exploiting large tire sensor data to build and deploy ML models via REST-API systems.
-
Communication of project status, milestones, and challenges to team members, managers, and
senior directors.
As a data scientist at the Luxembourg Institute of Science and Technology, LIST, I worked on applied research projects providing state-of-the-art solutions leveraging machine learning and multi-performance optimization tools in manufacturing for the automotive industry.
​
-
Providing input to the definition of research concepts in collaboration with LIST and Goodyear.
-
Design and develop a similarity algorithm to compare high-dimensional tabular data containing mixed-type (numerical and categorical) variables.
-
Developing kernel-based ML algorithms to predict multiple tire-related performances.
-
Developing a multi-performance optimization engine leveraging data-driven objectives and physical constraints.
-
Contributing to knowledge generation and dissemination of results as technical reports, research papers, presentations, and software packages.
As a Doctoral researcher at the University of Louvain in Belgium, I developed algorithms to extract knowledge from data modeled as networks, i.e., large-scale graphs. I contributed with new scientific methods applied to problems of anomaly detection on Amazon co-purchasing networks, predicting the toxicity of molecules represented as graphs, identification of biomarkers for predicting schizophrenia in the human brain network, among other projects, leveraging tools from machine learning and data mining.
Estimation of households' energy consumption from solar panels remains an important challenge in energy accompanies. My collaboration with NRB/Opinum Belgium consisted of developing time-series forecasting algorithms exploiting large amounts of meteorological and photovoltaic solar panel data.
