Most commonly discussed Machine Learning Algorithms are-
- Supervised algorithm which includes classification and regression algorithms.
- Unsupervised algorithm like clustering (k means).
Basically you need to understand first the taxonomy of machine learning .
Practically you have :
- Supervised learning : regression and classification task and basically understanding of gradient descent , logististc regression , performance metrics (rmsle , rmse, log loss , accuracy score) , and of course some ,learning’’ models that you could train on them such as regression , decission trees , random forest, boosting , bagging , adaboost etc.
- Unsupervised learning : see multiple methods for clustering
After that , neural networks , image recognition natural language processing .
It is kind of not realistic to know those alghoritms from scratch , because there are allready implementation for them , but you need to know what and when to use , eventual write ensembles over them.