Within the "Cognitive" business unit at Blue Reply, I worked on two different fields concerning Machine Learning models: explainability and fairness. Explaining a ML model is a process via which the predictions are made more understandable to humans by pointing out which features led to the final result. Fairness is monitored in machine learning models to avoid biased results, which instead might occur when, for instance, there has been a prejudice that has influenced the dataset used to build the model. I used different frameworks such as the proprietary one IBM OpenScale and different open source ones as Lime (by Marco Ribeiro), SHAP (by Scott Lundberg), DeepLIFT (by Avanti Shrikumar et al.), AI Fairness 360 (by IBM).

Tutor for architecture students of the course "Calculus I", an English course. I also prepared material for the course in LaTeX.

Tutor for engineering students of the courses Analysis I & II, both Italian and English versions.