1) Technical difference between Pyspark and Pandas 2) If you have large dataset with 1000s of features and only want some important feature, how will you identify it? 3) Dimensionality reduction? And why PCA is not an ideal solution in real world? Any other options? 4) If you have to build a model for Business partners regarding calls received by them for Opening account, paying bill and q&a? How will you design the whole ML algorithm from scratch? What features you will consider as important? How will you get the data? All steps involved till end. 5) For continuous and categorical features, which ML algorithm is best? And why?
