Candidatei-me online. O processo levou 1 semana. Fui entrevistado pela Aurora Solar em set. de 2021
Entrevista
The process was not heavy. All the interviewers are very friendly and they value your skills as well as whether you align with the company value. The scheduling was easy too.
Perguntas de entrevista [1]
Pergunta 1
Deep learning basic knowledge, a couple leetcode, and deep learning task
Candidatei-me online. O processo levou 3 semanas. Fui entrevistado pela Aurora Solar (San Francisco, CA) em fev. de 2022
Entrevista
Got a very quick response after my application to do a phone screen.
1. Phone Screen - standard phone screening, a mix of gauging your interest in the company, salary expectations, etc...
2. 1:1 - walkthrough with a manager or lead through the Aurora product. Asked some questions about past experiences.
3 . Technical Interview - The first part was a coding challenge, LeetCode style. I was not familiar but worked through it and did well. (I assume I did well since I was congratulated and told I did great. )
The second part was ML questions, trivia answer-like questions. I missed a few of them, which I believe led to my rejection later. (They said there were candidates with more experience) This is the biggest pain point for me: rather than relying on my problem-solving skills, they rely on trivia questions that have nothing to do with a person's competence. My advice is to drop questions like these, perhaps consider spending the time spent on trivia working through an actual CV problem to gauge a candidate's real skills.
4. There was to be a take-home project, but I did not make it to this stage.
5. I believe after that, there is a final interview with the team at large.
Overall the communication was really fast, and they had the courtesy to at least let me know my rejection soon after.
Perguntas de entrevista [1]
Pergunta 1
- Number all of the connected components in a matrix. - What is the complexity of the above function you wrote?
- What parts of the neural network take up the most memory?
- What is L1 Loss? What is L2 Loss?
- What loss can you use to help mitigate an imbalanced dataset?
(There were many more questions like this. Remarkably few were CV specific. For example, there were no questions on network architecture, CNNs, etc...)