I am working on an ML project. I have implemented it from scratch. I want to know what I am doing is right or not. Can anyone help me here?

That would be great!

I have shared the details

I'm working on a project where I need to predict if some customers are more likely to take loans if we provide them offers. It's a huge dataset with 100 columns. I have reduced it to 70 columns with 25lakhs rows. I tested correlation and done some feature engineering. I want to know about some of the steps. For example do I need to do feature engineering before train test split or after. Some of other before after things I want to know. I also googled about it but some suggest one thing others another. That's why I am in doubt.