Machine Learning Scientist Interview Questions
347
Machine Learning Scientist interview questions shared by candidates
The three data structure questions are: 1. the difference between linked list and array; 2. the difference between stack and queue; 3. describe hash table.
4 Answers↳
Arrays are more efficient for accessing elements , while linked list are better for inserting or deleting elements, the choice between the two data structure depends on the specific requirements of the problem being solved. Less
↳
Stack and queues have different order of processing, operations for adding and removing elements, and usage scenarios.The choice between the two data structure depends on the specific requirements of the problem being solved Less
↳
A hash table is a data structure that allows for efficient insertion, deletion, and lookup of key-value pairs. It is based on the idea of hashing, which involves mapping each key to a specific index in an array using a hash function. The hash function takes a key as input and returns a unique index in the array. In order to handle collisions (when two or more keys map to the same index), some form of collision resolution mechanism is used, such as separate chaining or open addressing. In separate chaining, each index in the array is a linked list, and each key-value pair is stored in a node in the corresponding linked list. When a collision occurs, the new key-value pair is added to the end of the linked list at the corresponding index. In open addressing, when a collision occurs, a different index in the array is searched for to store the new key-value pair. There are several techniques for open addressing, such as linear probing, quadratic probing, and double hashing. Hash tables have an average case time complexity of O(1) for insertion, deletion, and lookup operations, making them a highly efficient data structure for many applications, such as database indexing, caching, and compiler symbol tables. However, their worst-case time complexity can be as bad as O(n) in rare cases, such as when there are many collisions and the hash table needs to be resized. Less
How do you weight 9 marbles three times on a balance scale to select the heaviest one.
3 Answers↳
Do they all have the same weight except for one?
↳
I think the question is the rest of the 8 balls are the same weight. In that case you can find the heaviest in 2 tries. Less
↳
First weight 4 vs 4. If they're the same => the last one is the haviest. If not, weight 2 vs 2 of the haviest 4. Then weights 1 vs 1 of the haviest 2. Less
Fix a piece of sh*t code that just request data and save it to MySql. Super easy and easy to fix so I don't know what is the point of the interview. Totally a sh*t experience.
2 Answers↳
Can you please provide details of the technical round?
↳
Super easy, it didn't need a master student to do it. So I don't want to remember it since it makes no sense. Less
Given a number, how to find a closest number in a series of floating point data.
2 Answers↳
Use a Binary Tree data structure. It is sorted while it stores the elements. Then traverse the tree. Search is log N. In Java there is a ceiling and floor function which does a BFS traversal and tries to search the closest number rounded off, to the nearest below or above number on the basis of ceiling or Floor function. Less
↳
Just loop over the series. Calculate abs(ref_num-x_i). Remember index of last seen lowest abs difference. In the end just return the element at the index you have saved to the end. \theat(n) time \theta(1) space additionaly (besides space for series) Less
Probability, conventional machine learning, deep learning, coding/algorithm, behavioral
2 Answers↳
I was good with most of the questions
↳
Can you share what they asked in the coding round?
1- Machine Learning concepts 2- Data Structure Case Study (Very easy) 3- Coding (Easy but they need hands-on exp) And Finally, Leadership (14 principles)
2 Answers↳
There is no right answer except for conceptual one. They need you to be natural and be good in what you do. I would suggest to do following: 1- Leadership 2- Coding 3- ML concepts Less
↳
Forgot to add, my feedback was mixed saying I need more Machine Learning knowledge (which I disagree) and hands-on coding (which killed my interview)! Less
An online exam that takes about 2 hours. Most of the material is about deep learning and machine learning. Most of the answers can be found on the coursera class of machine learning and deep learning.
1 Answers↳
Hi, I have an interview test with deeplearning.ai. It says there's some video answering section at the end. any comments or suggestions for it? Less
Coding challenge prompt: Please send us an original code sample, the code should reflect the skills listed in the job requirements with base Python manipulations using the ML frameworks (Tensorflow, SKLearn, Keras, PyTorch).
1 Answers↳
We appreciate your feedback. Sorry you had a negative experience. We evaluate each code sample carefully and apologize for any delay as we’re a growing team handling large influxes of applications for all positions. Code quality is very important to us and we appreciate the effort and time you put into the application process. Less
we discussed mainly my background and some of the works I did, but she was not able to understand most of them. She was constantly interrupting me and trying to change the subject. The part that really got in her nerve was when I did not know what is 'original content". She explained that she means shows like "Orange is the new black", etc. which I do not watch. It was completely incomprehensible for her that some people may spend their time reading papers, or help their church high school kids instead watching fictional stories about women in prison.
1 Answers↳
If you want to pass the initial phone screen, make sure you know what is "original content" ! Less
What are the assumptions on independent variables for a linear model? Is the p-value valid if the independent variables are not normally distributed?
1 Answers↳
There is no assumption on the distribution of independent variables. Residuals have to follow specific distributions for p-value to be valid. Otherwise Monte Carlo simulation can be used to calculate p-value based on the sampling distribution of the parameters (interviewer was convinced that the answer is wrong and that normality of independent variables is necessary. I wonder how he treats 'dummy variables' that take values 0 or 1 as 'normally distributed') Less