0. Self-introduction 1. Project introduction, focus on performance optimization 2. Short-answer quiz on GPU architecture, CUDA programming, synchronization methods, etc 3. Oral CUDA programming, matrix transpose, parallel reduction 3-1. Performance problems, memory access patterns, warp divergence, and so on 3-2. Optimization 4. Question on C++

It was telephone screening which lasted 1 hour with different questions in the area of computer architecture. The interviewer was very calm and gave sufficient time and hints to answer. First was contacted by recruiter who setup an webex interview and the whole process tooks less than a week.