Machine learning research scientist Interview Questions

26

Machine Learning Research Scientist interview questions shared by candidates

Top Interview Questions

Sort: Relevance|Popular|Date
Bloomberg L.P.
Research Scientist/Machine Learning was asked...November 27, 2014

More on machine learning algorithms. Easy coding question

1 Answers

Can you provide more details on the questions please?

Amazon

Describe a project where things didn't work out as expected.

1 Answers

Gave example

Amazon

You have a linked list of numbers, how would you return the median ? Follow up, what is the worst case performance?

1 Answers

Basic data structures. Sorting algorithm. Calculate indices used to return or compute the median. Use indices in loop. Less

Bloomberg L.P.

After you have decided which features to use, describe the process of constructing feature-vectors

1 Answers

It was almost all about tf-idf vectorizer

Amazon

You have a singly linked list, how would you find the median?

1 Answers

Sort it (choose an algorithm), calculate index for median, loop over next based on index. They will want to actually discuss implementation (i.e. use a for loop, traverse next ptr, etc.). You should also know algorithmic complexity (big-O notation). Less

Amazon

Describe a project where there were multiple alternatives from which to choose in implementing it.

1 Answers

Gave examples from work history.

Rocket Fuel

'''Question 1: Given a sorted but rotated array, and a target, find the location of the target in the array. If the target is not in the array, returns -1 1) INPUT: [3,6,7,1,2], target = 1 OUTPUT: 3 2) INPUT: [3,6,7,1,2], target = 9 OUTPUT: -1 '''

1 Answers

def search(A, target): if target == None: return -1 #if not target: # return -1 #riht I needed to address this! try: target = int(target) except ValueError: return -1 l=0 n = len(A) r = n-1 while l<=r: middle=l+((r-l)/2) if A[middle] == target: return middle if A[l] <= A[middle]: if ((A[l]<= target) and (target Less

Amazon

If you were designing a robot lawnmower, where would you start?

1 Answers

They will ask you about how specifically you implement the perception algorithms, what sensors you will use, and where you put the sensors. They were more interested in computer vision. Less

Bloomberg L.P.

How do you implement feature selection in text classification?

1 Answers

Mutual information / chi-squared criteria

Amazon

If you were designing a robot lawn mower, where would you start? Various follow ups including sensors, limited to cameras, where mounted? How would you recognize people? Describe skeletal model?

1 Answers

This was basically describe the technical solution and they wanted also to know specifically how I would implement them. Less

Viewing 1 - 10 of 26 Interview Questions