2024 - today: Building machine learning pipelines
Place of Work: Remote
Role: Data scientist - Data Visualization Expert
Customer: CTGLab, Italian National Research Council (CNR)

Tasks:

• Building machine learning pipelines in tidymodels to enhance Colorectal Cancer diagnosis by analyzing patients' biomarkers from urine (including microbiome data) and from blood. After model type selection, feature selection and hypeparameters optimization via grid search on the training set, the models achieved high predictive performances on the test datasets (AUC ? 0.9)
• Co-authoring a manuscript on the ML-based diagnostic pipelines for Colorectal Cancer, targeting submission to a peer-reviewed journal in computational medicine
• Developing and deploying ReViewCNV, available on GitHub as either an R package or a Dockerized ShinyApp for visualizing copy number variations (CNVs) in patients' genomes. The tool aids clinical interpretation by providing genomic context, showing copy number polymorphisms in correspondence of identified CNVs, and allowing family analysis
• Collaborated with interdisciplinary teams (bioinformaticians, clinicians, developers) to deploy reproducible analysis pipelines and user-facing tools
• Using standard practices (Roxygen2, testthat, GitHub actions, containerized deployment) to improve reproducibility and reusability of custom analyses

2013 - 2022: Teaching
Place of Work: Milan - Italy
Role: Natural Science Teacher
Customer: Scientific High School ?Guglielmo Marconi?

Tasks:

Gained deep experience translating complex scientific concepts to non-expert audiences ? a skill now applied to data storytelling and communication in bioinformatics