Company: Amazon India
Location: Bangalore
Position: Applied Science Intern (Machine Learning)
There were 2 interviews in consecutive days.

I am a dual degree (B.Tech and M.Tech) student and currently, I am in my 4th year (pre-final). I actually got a referral through one of my professors with whom, I am doing my masters thesis project.

1st Interview:
This was on data structures and algorithms. Questions were fairly simple. I would say 2 leetcode easy and 1 leetcode medium. Cant remember the exact questions, but they were based on heaps, string, and hashmaps. I got all of those questions correct. They were really straightforward. Since this is a ML position, it might be the case that people arenot asked difficult DS questions. I might be wrong and lucky though.

2nd Interview:
This started with my introduction where I told some of my ML projects in brief, my interest apart from Machine Learning, that is backend Web Development. The interviewer was really grilling me on the fine details of my project, like, what architecture did I chose, or if I used transfer learning or not, what went wrong. Basically, any term I throw in, he was asking me to explain it. This took almost half hour.
Then he asked some ML concepts such as:

1. What is bagging and boosting? What is Adaboost. How is it different from Adagrad
2. Transfer learning.
3. Some maths related to batch normalisation. Is normalisation done on whole dataset at a time or a just the current batch?
4. Explain resnet architecture. (I explained it with great length). Then he keep asking about some other architectures such as EfficientNet, MobileNet. I didnt knew how they operate, so told him honestly about it. Basically he kept grilling on a particular concept until I give up :/
5. Decision Trees and Random Forest. (Amazon particularly loooooove random forests 😂) How to split the dataset, whats the objective (like, decrease in entropy or gini)
6. Maximum Likelihood Estimation. How is it related to logistic regression.

I didn't had any HR round.

So overall, the experience of interview was really good and relaxed. With Amazon, I had one of the fastest interview, to be honest. The HR was really helpful in setting up the interviews and after 1 hour of interview, I got her call about my offer.
So, my advice for anyone who is going for a ML internship: Just be honest and rigourous about your work and concepts.

Cheers 😃