hi I am making neural networks to solve supervised/unsupervised problems, I can resume the process to show you how much I know in the fild.
first thing we should have dataset cleaned and prepared, in case we haven't a dataset we can make use of web scraping to collect data, or if we have small datasets some cases we use augmentation techniques.
the last thing we should do in preparation step is to split dataset into three steps train/validation/test sets.
the second thing is build and train a model so we measure the accuracy and we make some analysis to figure out if occur overfiting or other problems we the results are good we can move on and deploy our model if necessary using flask with Python, also I have knowledge on aws ec2 instances if we want to deploy it on the cloud