I won't divulge the custom questions they had because of an NDA. But I will divulge the interview questions that are carbon copy menial tasks. Of the first three phone/live-coding interviews, I was asked to do a merge sort. I guess I understand that they want to make sure I can code. But there's got to be a better problem. The many in-person interviews were (1) 1.5 hour coding-by-myself on a laptop, (2) describe a problem you worked on in detail, (3) "NLP" interview that didn't focus on NLP, instead they focused on AB testing (which is fine), (4) Lunch interview, (5) culture fit interview, (6) follow up/close/discussion interview. To be honest, I don't remember much near the end. I think I went directly home after and slept for 12 hours.
Nlp Research Intern Interview Questions
125 nlp research intern interview questions shared by candidates
write code in python with DSA
How to disambiguate a/the in a text.
Suppose we are trying to build a text/topic classifier of Twitter. The classifier is trained offline. When we put it online, how to deal with unseen data?
Describe how to use HMMs to build such a classifier.
How to you deal with project difficulties?
What is the difference between dependency parsing and constituency parsing?
First screening interview: Q1: About background and projects Q2: If you are given an encoded string, design the algorithm to learn decoding scheme and decoded string. Design Machine Learning or Deep Learning approach or any other method. Q3: Some questions on basic Machine and Deep Learning. Some questions related to analytical thinking. Virtual On-site interview: Q1: implement (unfold) simple tokenizer and BERT tokenizer code. Q2: Implement the Beam Search algorithm at a very high level when you already have decode for Machine Translation (say English to Spanish).
Not difficult but that was what I was asked: I was asked to write an algorithm and implement it to detect whether a search engine query is a question or not.
Basic Python questions, Projects from resume, BERT model, knowledge on transformers etc.
Viewing 11 - 20 interview questions