I like designing crossword puzzles, every now and then. So, I coded a simple crossword playground for whoever would enjoy solving them.

With a colleague at The University of Bologna, we designed a basic iOS app for hand gestures recognition. We used MobileNetV2 and retrained it on two different datasets, one for "numbers" (coloured input) and one for "gestures" (greyscale input). Then, we converted and merged both trained Core ML Models in one iOS application.

With a colleague at The University of Bologna, we participated at the Tablut Competition for the AI course that is part of the Master's Programme we're attending. We implemented a client for playing Tablut (an ancient board game, Ashton version) that chooses the best move to perform on the board, then updates the state and sends it to a server. The move is chosen using classic Minimax AI algorithm with Alpha-Beta pruning and a search depth of two.

To enhance my crosswords experience, I thought it would be great to have them playable in a mobile app. So, I coded a cross-platform app that can be built both for Android and iOS. The project was mainly coded in Dart language using Flutter VS Code Extension.