The current state of Deepfakes and PyTorch - some guidance please

I work for a large company where we are trying to constantly promote awareness of security threats, things like phishing attacks, ransomeware, deep fakes, etc.

I have been asked to give a demonstration on deep fakes, where I will use some public/free/no copyright material and deep fake it with someones face from our organization.

Does it make sense these days to do build this in PyTorch, or is there now such good off the shelf software I should just use that? Its a basic short demo, 10-30 seconds. I don’t want to put more work into than necessary. My company, and myself have no interest in Deepfake technology other than this demonstration on security awareness, which we would like to make an original deep fake.

If PyTorch is a good approach and anyone has some good links, starter code, ideas, etc to share please post in this thread. If a commercial app is a good solution please share those. I am currently researching both. I am quite comfortable with PyTorch and Deep Neural Networks, but obviously I am trying to figure out the approach with least effort and good result. Thank you for any help and guidance.