what's up everybody this is Jay from interview query and the data science today youtube channel and today i am talking about the microsoft data science super exciting also before i get a chance to talk any further that love if you guys could just like this video right here really helps with encouraging me to make more of these and I don't get enough encouragement in every single day so please like it what is the interview like Microsoft right I would say Microsoft in general has changed over the past few years better well known as decade right ever since the new CEO came in I think the culture has definitely changed into something more of being more effective in terms of business priorities fairing out where where they actually lay on the SAS kind of roadmap and generally become one of the biggest players in data science and machine learning with applying you know basically ml to I think every aspect of their b2b huge multi conglomerate business and that comes from everything in artificial intelligence and a juror to actually data science consulting for their cloud computing products it's humongous right Microsoft has a huge data science at work so how does the actual interview go right so I want to start out by kind of explaining how data science functions at Microsoft essentially data science is a byproduct of under the engineering department and it functions within three different roles right there is the actual pure data science role there is an applied scientist role and then there is the machine learning engineer at all right and these are the three roles that effectively make up all of the data scientists at Microsoft and effectively depending on their function they work in a variety of capacity but mostly the machine learning engineers and applied scientists work on writing code to ship models to production writing code to use machine our learning algorithms that can be used by other data scientists on the team a lot of the data scientists will work with customers directly or indirectly to solve technical issues work on metrics and experimentation and also focus on product features right so there's a definitely a distinct difference between the applied scientists and machine learning engineers versus the actual data scientists that focus more on product facing capacities so what does the actual interview look like then right from the research that we've done an interview query and talking to current and past members we see that mostly the experience consists of an initial phone screen and then a technical interview which then concludes with an actual on site now kind of virtual interview with five to six different different members of the team right the initial screen is generally just a recruiter screen and it kind of depends on actually if you're a more senior than a hiring manager will actually go out and talk to you about the data science role and try to understand your past experience I would assume that generally this is pretty straightforward just like most other interviews that you'll ever have in that the hiring manager is just trying to get a sense of your background and see if you're a good fit for the team and Microsoft as a whole right they want to see if your resume matches your experience then at the same time they might ask a few technical interview questions as well but more around the concepts of data science right so this is something like explain the difference between lasso and Ridge regression right or how would you actually explain a deep learning model to a business person these are more questions that are dependent upon how you can communicate technical concepts to non-technical persons as well as can do you understand the fundamentals of machine learning algorithms from a level that is deeper than just you know applying scikit-learn to something right so this is more about the fundamentals and I think this comes up many times for Microsoft type of interviews in that they want you to be generally I think they have preferences towards people that come from some sort of academic setting but also someone that has a deep understanding of data science and NL as well and isn't just there to just you know apply you know libraries to everything etc right so after that generally there is a technical screen with the Microsoft data scientist and this is usually 45 minutes to an hour and it consists of a couple of different questions from sequel probability statistics to algorithms expect a question within Python that is going over anything as simple as just effectively you know parsing a string to something a little bit more complicated as a kind of more of a leak code style algorithm question right this definitely depends on your team and depends on how technical the role is right so the more technical the role I expect the harder the question then they'll usually ask like some sort of sequel question or a statistics and probability question where they're asking you effectively do you know like the fundamentals of probability in stats do you know the fundamentals of sequel yes so then that's good cool now lastly is the on-site interview right and this is a full-day event usually it goes from like 9 p. m. to like 9 a.
m. sorry to 5 p. m.
