Candidatei-me online. O processo levou 4 semanas. Fui entrevistado pela Shopee (Singapura) em abr. de 2020
Entrevista
I want to preface this review by saying Shopee is a fast growing company, where processess and policies can change quickly. What that means is an agile HR system that learns from itself. Compared to the Shopee of two years ago that I previously interviewed they've revamped the whole process entirely and made it much smoother for engineers overall.
I interviewed with Shopee ML team in late february / march. They quickly sent me an online assessment, which requires you camera, microphone and screen share. Afterwards the recruiters set up a quick call to touch base and scheduled me to do a phone screen with another engineer. This interview had a small coding portion and then a series of quick questions related to ML concepts. I wanted to highlight the depth that my second engineer screen went into to check my knowledge of data structures.
Afterwards I scheduled another similar phone screen a week afterwards, then a call with the hiring manager of the team. The two engineer screens had a lot of overlap in terms of content. The hiring manager screen could be improve a little if his questions had more structure to them and or if he asked more about my working style and how i would fit in with the team. Some questions about my latest favourite papers and some more things in that direction may also be useful.
In particular I want to highlight my recruiters Ronda, Andrea and Fion whose professionalism and responsiveness really made me think positively of Shopee. Their ability and answer my questions and give me quick feedback made the whole process enjoyable and it was really a pleasure to work with them.
Perguntas de entrevista [3]
Pergunta 1
Some questions to test your understanding of ML applications and approaches
Went straight into coding problem once the video call started, without any introduction or other converstion. Typical technical interview with coding problems of leetcode hard level. Also asked some machine learning conceptual questions.
Perguntas de entrevista [1]
Pergunta 1
How do you calculate AUC given a list of positive and negative instances sorted by the predicted score?
It was ok. There was a coding challenge followed by 2 interviews. Coding challenge was doable. 2 interviews lasted 1.5 hours each. One peculiar thing that happened was during the interview process they switched the exact team i was interviewing for 2 times.
Perguntas de entrevista [1]
Pergunta 1
Technical interview asked about performing matrix multiplication in an efficient manner and a couple of questions regarding machine learning
1. HR phone call to decide which team you want to interview with. Also asked about salary expectations.
2. Technical interview on Zoom with an engineer
3. Technical interview on Zoom with the team lead
Perguntas de entrevista [1]
Pergunta 1
The HR call was very short and I did not get asked behavioural questions. The HRs are very nice and reply to emails very quickly.
Then the technical interviews. It seems that they have a pool of questions that are shared with all engineers. So the first interviewer asked me backend questions even though I applied for MLE. The interviewer was nice about it when I told them that I have not learned certain topics in school.
I was asked about the Machine Learning projects on my resume. They wanted to see how well you know the specific concepts and ML algorithms you used in these projects.
For the coding challenge, I got a pretty easy question in the first round, about designing a cache. The question asked in the second round was quite difficult and required dynamic programming.
Also, even though the interview started in Mandarin, the interviewer switched to English when I told them that I did not know technical terms in Chinese. We had no problem in terms of communication.