Answer 1

python
https://www.pyimagesearch.com/2018/06/18/face-recognition-with-opencv-python-and-deep-learning/

Answer 2

I am still learning the basics behind Machine Learning (including a lot of the "fascinating" math). But if you are okay with not understanding the inner workings of the libraries you are using, you can just do a short cut and start with one of the tutorials you can find online like this one:
https://www.pyimagesearch.com/2018/09/24/opencv-face-recognition/
As far as I understand it OpenCv is great for preprocessing. It uses deep neural network libraries like caffe and torch to extract the relevant features from images. To actually train a model (to find the parameters to rightly classify new data) you need to decide which ML-Algorithm or Neural Network you want to use. In the course linked above a SVM (support vector machine) is used. But openCV also supports neural networks frameworks . So for the training of the parameters you can use frameworks like sk-learning (for more traditional machine learning with SVM, kNN, randomForest etc.), pytorch or tensorflow. But please don't ask me for any specifics. I am not there ... yet.

Answer 3

Yeah, the little cart is sometimes enough to get things done but at other times you need that van. ;-)