how to prepare Ml case study?

how do we prepare ML case study ? from the interview, i did follows: 1. define which data i need(features) 2. pick few simple models(RF, SVM etc) 3. explain how to prevent overfitting, 5. move into DL algorithm if they ask but i think this was a not enough shot since i was rejected to Bloomberg nlp, google Ml. i always pass the phone interview(Js, Twosig, Amzn, google, bloom,lyft etc) since most of them ask theory or stats. but i dont pass on site whenever they ask case study. any suggestion will be very appreciated! here we go for blind commitment(TC ..) company: Faang TC 195k yoe 1