As @YCR correctly said, decision tree is part of machine learning. The answer is "it depends". It depends on the data type, quantity, and on the computational resources that you have for the training. For tabular data sets, a very good starting point is random forest (ensemble of decision trees). You can find implementations in many libraries like sklearn and opencv.

If you have an image database, like photos that should be classified as jobs (a painter, a farmer, and so on) I recommend convolutional neural networks like ResNets. A nice trick to avoid a lot of computational time in training is using transfer learning (for example from the ImageNet database).