Just to add that he has a video series on DL (youtube), completely approachable and accompanied by code notebooks.

How does it compare with Andrej Karpathy’s video series on building GPTs from scratch? Are they pretty much teaching the same things?

Karpathy focuses on GPT, well, NLP-related specifics, while Raschka overviews Deep Learning as a whole, starting from the Perceptron basically.

Karpathy's teaching style is well, Karpathy, Raschka is more conventional (but not buttoned down).