• Lead the creation and conduction of product demos, during the pre-sales phase and public events, showcasing our solution's capabilities to prospective clients• Provide technical support throughout the sales process, identifying clients' business needs and advising on the best product implementation and integration with existing IT infrastructures and leading databases and data warehouses• Conduct training sessions for end-users and IT staff, ensuring they are fully equipped to utilize our product, and provide continuous support to maximize customer satisfaction

• Conducted Presales technical presentations and Product Demos alongside the Sales team, interacting closely with customers to determine their specific needs and objectives;• Supervised and mentored a team of Data Scientists and Software Engineers, to design and deliver innovative solutions that met both customer requirements and company goals;• Involved in entering into a strategic partnership with Landing AI, a leading American company specialized in Computer Vision applied to industrial field; exhibited LandingLens demos during important tech fairs, like MECPSE and SPS;• Contributed to Addfor's flagship Product launch, Nutray, in the Italian and European markets, while managing the product backlogs and roadmap; worked with external design and production experts to create the hardware and tailor the product to customer requirements.

Title: Sanity Checks for Explanations of Deep Neural Networks PredictionsSummary: The lack of transparency of Deep Learning models represents a barrier to the adoption of these systems for those tasks where interpretability is essential. To address this issue, the Explainable Artificial Intelligence (XAI) field has developed a suite of methods and algorithms for visualizing, explaining, and interpreting deep neural networks predictions. However, the difficulty remains in assessing the scope and quality of the results provided by these explanation methods. The goal of this thesis is to validate the explanations of deep neural network predictions for Computer Vision, generated by state-of-the-art methods, and assess their reliability by performing experiments to test their sensitivity to model parameters. The analysis conducted aims to provide useful insights into developing better and more reliable visualization methods for deep neural networks to gain the trust of skeptical users in high-stakes tasks where visual inspection of results is not easy.