🤝 My name is Giovanni del Vecchio and I live in Milan, Italy.
🎓 I moved here to study at Politecnico di Milano where I got my Bachelor's degree in Computer Engineering. Currently I am working on my thesis project that is the last step to take for my MSc graduation in Computer Science and Engineering.
👓 My areas of interest reside mainly in Machine Learning and Data Science topics, but I am always eager to learn new technologies from different fields.
🏉 I love doing sports and I am currently a Rugby player

• Artificial Neural Networks and Deep Learning Project:
  The project consisted of three subtasks:

1. Image classification: about 95% accuracy reached in classifying images with or without people wearing masks by using a fine tuned CNN and transfer learning from well known ANN architecture.
2. Image segmentation: accomplished to reach a result of more than 0.70 in IoU metric by building a U-Net-like ANN exploiting tiling and image augmentation techniques.
3. Visual Question Answering: about 60% accuracy reached in answering questions related to given images using LSTMs and Co-attention mechanism.
   TensorFlow Keras NumPy Pandas NLP

• Recommender Systems Project:
  Design of a Recommender System for an online book shop combining various kind of systems like: User based, Item based, Matrix Factorization, SLIM, Factorization Machines, Graph based. Exploiting some Bagging and Boosting techniques (like XGBoost) our Recommender System reached a MAP@10 of 0.1015.
  NumPy SciPy Pandas MatPlotLib Cython Matrix Factorization
• Data Intelligence Applications Project:
  Accomplished to yield the optimal combinations of advertising cost per click and pricing strategies in a virtual scenario assessing the goal through CTR and conversion rate metrics, for a discrete set of customers categories, by tuning and combining different online Multi Armed Bandit algorithms such as UCB-like bandits, Thompson sampling bandits and Gaussian Processes bandits.
  Scikit-learn MatPlotLib NumPy MAB
• Software Engineering Project:
  Digitalization of a boardgame called Adrenalina in Java language. Implementation through Object Oriented Programming classical patterns like MVC. Design of multiplayer mode using Socket and RMI for client/server communication. Built two different User Interfaces: one implemented with JavaFX and the the other in a CLI style.
  Java OOP MVC client/server UI