I applied online. The process took 4 weeks. I interviewed at Magic Leap (New York, NY)
Interview
The first round was a phone call with the recruiter. Next, was a 45-minute video interview with an engineer. He was dumb af. He constantly kept asking me the same question about my research and I explained it to him every time he asked. I also referenced my published paper which has the details of the implementation. He wasted 35 minutes into convincing me that my published paper research is completely wrong because he was unable to comprehend, and asked me extremely redundant questions related to the published paper of my research.
Finally, in the last 10 minutes of the call gave me a lengthy coding question of the leetcode level hard. He expected me to read + understand + code in 10minutes. This is unfair.
After two days I get an email that I didn't make it to the next level.
I still fail to understand what was the interviewer trying to accomplish. He was completely unfair.
HR talk, introduce the company, one video screening with technical knowledge and coding question, one online assessment with harderrack, then another video screening talking about machine learning, then onsite
Interview questions [1]
Question 1
epipolar geometry and machine learning optimization
based on resume
I applied online. The process took 4 weeks. I interviewed at Magic Leap (Sunnyvale, CA) in Jan 2019
Interview
The phone screen ask about past experience and related projects. The manager introduced the company and what they are currently doing. Then there are some questions about case studies about real-time reconstruction and recognition stuff.
After a week, the HR quickly responded with an onsite invitation. The onsite is composed of five rounds, first a representation for previous works. Then some coding+research questions, probably 50/50, are asked by different reviewers. Most of them have Ph.D. degree.
Interview questions [1]
Question 1
The coding questions are quite simple. Research questions are about multi-view geometry, such as fundamental/essential matrix and deep learning basics (resnet, loss function, kernel size, pooling, dropout, batch norm, etc.). Generally the questions are not so difficult.
