I am a python developer and want to create a app where i can detect and give suggestions to users about their posture on various points while performing a activity. (i.e. Playing a specific game) I know the rules of game like how to stand and what to do. I want to develop a app where user will upload a video and they can get suggestion at different point of playing game. Video will be about 5-9 seconds. How to approach this feature and how to make this happen. I have no experience with machine learning.

I have tried to get coordinates of body parts with mediapipe and tried to analyze. But i am not sure if this is correct way to approach this problem. It would be great for any suggestions on how to implement this what to use etc.