My advice would be to start with practical projects and then slowly progress with theory.

Kaggle notebooks are great to learn the practical part: https://www.kaggle.com/notebooks.

Ask questions in https://www.reddit.com/r/MLQuestions/ .

When you become satisfied with your knowledge of tools and practices, I would advise that you construct the dataset for some problem by yourself (eg. you can scrape the data) and apply ML algorithms to it. The hardest thing in ML is dataset construction. You might even build a company out of it😊