Status: MS Computer science
Position: Data science intern at Amazon
Location: Seattle, WA
Date: June 03, 2019
I applied for a data science/machine learning fall internship opportunity at Amazon in May 2019. I was contacted by a recruiter on Jun 3 and she scheduled a 2 interview. Everything went very fast...
The first interview was on june 11 and the next was june 12. The next day(june 13), i was contacted by the same recruiter for a 3rd phone interview which was schedule for Tuesday june 18. I got an offer on June 25 : ).
Preparation Prior to Interview:
I am a computer science Masters student with a background in mathematics. I am in my second year and i have no programming experience before, nor had i ever heard about data science or Machine learning. I took a bunch of online courses in python, data science, statistics, Machine learning, SQL. I am a competitor on Kaggle whichhelps hone my data science/machine learning skills, I practice coding with python on leetcode and Hackerank, I also practice SQL on Hackerrank. Before interviewing for amazon, I had a bunch of botched interviews with other big tech companies and i think that also count as interview practice.
After being scheduled for an interview with amazon, i solved all the easy top amazon questions on leetcode, about 60% of the medium questions(remeber to always check their time and space complextity and try to come up with a solution that has the best of each). I solved a few hard questions on leetcode and a lot of data sctructure questions too(my goal was to do atleast 5 codes everyday).
I revised all the machine learning models i have every used on my resume, brushed up on some past data science interview questions and answers(can finds them on glassdoor, springboard, youtube...).
Phone interview 1: 1 hour using the amazon livecode
- Short intro from the interviewer
- Resume/Behavioural questions
- Tell me a little about you self
- Looking at my resume, Tell me about the best project you have worked on and why is it your best project? (it was a team project so the next question...)
- How was it like working with a team and what was your part?
- What were some of the difficulties you encountered while working on that project with your team members?
- Is there any thing you would have done better or any future plans for this project
My advice for this section:
- JUST RELAX AND ANSWER THE QUESTION IN A CONVERSATIONAL MANNER.
- Have a copy of you resume handy and only talk about the project you know best
Technical/coding questions:
- What course have you taken? What algorithms/data structures did you learn?
- What are the different sorting algorithms you know?
- Explain how a quicksort work. What is the time complexity of a quicksort?
Coding:
Given a string, find the repeating characters in a string and return the number of time they appear. Leetcode question easy. What is yout time complexity?... can you write a more efficient code?.... what is the time complecity.
Last 5-10 minutes: Any Questions you have for your interviewer.
Phone interview 2: 1 hour on the phone
Mostly behavioural and conversational but a little stressfull through the end:
- Tell me about yourself
- Why did you transition from maths to computer science?
- How did you get into machine learning/data science
- Looking at you resume, if i gave you a scissors to cut out a 2 inch section in your resume to keep, which part would you keep? why?
Now the stressful part:
Suppose you work at Amazon and you want to understand why customers in region A have a lower review rating(70%) as compared to customers in region B. What are some of the question you would ask yourself to justify this ratings.
Yeah!!! not very clear right ? I had to ask a lot of questions, try many different answers, took me some time to get the "right" reasoning, while he was constantly pushing me to answer the question, think hard at the answer, repeating the question....
My advice on this section:
- Do let the stress get the best of you. I think it was the interviewers way of accessing how well i can handle and reason through a stressfull situation.
- Ask as many questions as possible, until you understand what the interviewer want from you
Last 5-10 minutes: Any Questions you have for your interviewer.
Interview 3: Video Phone call using Amazon CHime app
brief introduction of the interviewer
- Tell me a little about yourself
- walk me through the different parts of your project- data cleaning, replacing missing values, checking for skewness, creating trfaining, validation sets...
- What are the different preprocessing algorithms you used in your project(specific project on my resume).
- Why did you use this performance metric(f1-score in this case) instead of Accuarcy?
- What other classification performance metrics do you know( define them;f1-score, precision, recall,...)
- Why was it important for you to standardize you data set and what transformations did you use(used np log transformation).
- Why did you have to class oversampling on this data set. Did you do the oversampling before or after splitting you data set into train and validation?
- what are differet algorithms you used in this project and can you tell me a little about each of them.
- What is ensembling? Have you ever used an ensembling algorithm
Coding 2:
Reverse words in a string( leetcode easy) :'). What is the time complexity?( https://leetcode.com/problems/reverse-words-in-a-string-iii/)
Last 5-10 minutes: Any Questions you have for your interviewer.
I hope this is helpful. Goodluck on you interview practice.