Round 1 - Online Test There were 8 Questions consist of (3 single correct and 3 multiple correct Questions) Mcqs based on Math(Rank of a matrix,type of matrix etc),SQL, statistics distribution, and ML fundamentals . 1 question based on converting given dataframe to output dataframe , following by applying options like: drop all the row where column has >3 value clear a specific column value into alpha-numeric characters (regex use karke hogaya) and basic data cleaning wo sab Last question was on SQL query (medium level leetcode) (Using joins, order by ,count ,group by) 12 out of 22 were selected for next round Technical Round 1 (1 hr): Introduce yourself He asked me how familiar are you with data structures,then one easy level question Find whether an array is a subset of another array, I explained multiple solutions and asked me the complexities of all solutions which. Then he moved on to machine learning. He asked me each and every algorithm, their uses ,when to use what, and the advantages of one algorithm over another. For eg Logistic Regression and Linear Regression. Then asked some internship related questions,what you did there , what was your role. After that he asked about neural networks.what is dropout , how it actually works,why it is used,you need to use exact terminologies for every answer ,you cannot give vague answers. What matrix would you use to accuracy(precision and recall) what is precision and recall and formula What is batch normalisation,when to use Then asked me different activation function ,why to use activation function Then what is relu ,formula ,problem with relu and which function is used to overcome the problem(leakyRelu). Difference between list and tuple Any questions for me?2 -3 qs puch lena 7 out of 12 were selected for next round Technical Round 2 ( 45 to 50 mins): Totally Application based(should have thorough knowledge of ML and DL) Introduce yourself What challenges you faced in your internships, projects , what are the uses of the projects Then he deeply asked me about algorithms i used in my project , how they work , what is maths behind the algorithm , why to use , why you choose to use this algorithm only, he asked about all the projects i mentioned in my resume and grilled me on all the topics Again asked me precision and recall, formula. Suppose there are 100 data points out of which 90 are positive and 10 are negative ,what is the accuracy of positive data points. What is loss function and cost function If the cost function is y = wx+c then what is the loss function for that What is decision tree and random forest , when to use what, what is the difference. What is bagging and boosting, eg of bagging and boosting. Difference between linear and logistic , are they multivariate or bi-variate What is svm What is object detection , how to calculate accuracy of object detection , which model to use Any questions for me?2 -3 qs puch lena Results: Selected!! 5 out of 7 were selected ! Tips: Be confident and show you are very interested for the role . Be honest agar nahi pata to sidha na boldo,they are not expecting us to answer all the things. Don’t give vague answers , use proper terminologies Just Be yourself.

First Round: 6 mcqs with stats questions One python programming question One Sql question Technical Interviews: 2nd round: ML breadth/depth( 45mins) Logistic reg, Decision trees, CV, NLP, formulas, concepts, mathematical explanations, Tensorflow, Pytorch 3rd round: SQL focused (30 min)