In mgr_thesis, there is a jupyter notebook with a variational autoencoder I was working on recently.
In racetrack_game, there is an implementation of the square paper game Racetrack (with an AI opponent using a combination of A* and greedy search algorithms).
In deep_learning_assignments, there is a few solutions to some tasks from the Deep Learning course on MFF UK I passed.
In intro_to_ML_assignments there is a few solutions to assignments from the Introduction to Machine Learning course on MFF UK I passed (the code there is not as nice as in some newer tasks, but it implements Naive Bayes, SVM, Kernel linear regression, etc. in numpy, which I thought would be interesting).