Thanks for watching! Just so you know, we recently launched our new data science learning path - the most comprehensive we have so far! It features more than 30-course modules and will teach you how to answer every type of data science interview question. Try it here: www.interviewquery.com/learning-paths/data-science
Thanks. I have DS interview in 1 week. Previously, I had questions for MLE role: 1) Vanishing or exploding gradients - intuition behind the problem 2) Difference between random forest and decision tree and XGboost? 3) difference between word2vec and BERT 4) describe when the result may have statistical significance 5) difference between tf idf and count vectorizer 6) what is semi supervised learning and my experience with it and etc
I am interviewing for. Data scientist, applied scientist &. mle roles. I saw some lovely questions like :- 1) How are CNN & LSTM's related? 2) Structure of LSTM and full BPTT through an LSTM. like i was expected to describe all backprop equations 3) Can you use CNN for time series? If yes then what's the architecture? 4) Pseudocode of GBDT and how will it change if base learners are not shallow trees? .....and so on I rarely see business case, product metrics or A/B testing questions. Probably because I'm a fresher.
@@iqjayfeng I think these questions depends a lot on the background as well. I come from Maths background (honours) so these questions are expected (like I was literally asked to do derivation once in an interview). The third question was asked for Applied scientist @ Google. Also 'hardness' is subjective as I can rarely answer your business case and product metrics questions.
as i start applying to ft roles, i've been seeing ml engineers or ml swe popping up more frequently. Do you think you could makes videos on the diff roles between ml/ds? I've also been reading that the ds career track has a high floor but relatively lower salary ceiling. Idk what to consider and it might be insightful for others to hear about the diff options!
