Everything that is done with python uses the same virtual env

Nothing is better than code, here is the official repo by Microsoft with examples

The textual documentation can be found here: https://learn.microsoft.com/en-us/semantic-kernel/ai-orchestration/kernel/adding-services?tabs=python

Create a local environment with python, then allow the ipykernel to create a kernel for jupyter notebooks.

python3 -m venv skernel
source skernel/bin/activate
# For windows you must use the following command to activate the virtual environment
#  .\skernel\Scripts\activate 

you can handle requirements with easy thanks to pip

pip install -r requirements.txt
pip freeze > requirements.txt

Then you can create a kernel for jupyter notebooks using the very same environmnent

pip install ipykernel
python -m ipykernel install --user --name=skernel_experiments

Kernel can be removed using

jupyter kernelspec remove skernel_experiments