I recently completed the course and then wrote a script to automate the training and prediction process for random forests. The script also automates the feature engineering process and parameter ( for the random forest like min_samples_leaf) choosing process. The link to the blog can be found here :
Nice work! and I would recommend wrapping your script in a Class (you should be able to do so if you implemented the random forest from scratch ) and try to break down the big functions into smaller ones to improve the readability of your code.