The process started with a brief meeting with Technology Recruiter, he was a nice guy and explained everything in detail. First stage was a 30 mins chat with Lead Data Scientist-Computer Vision. The person asked about my previous experience, the projects I worked on and the choice of algorithm along with reasoning. The second stage was take home assignment, with a time limit of 1 week. Difficulty level was easy-medium. Those were superstore environment related Computer Visions Coding tasks. They rejected my application with a feedback criticizing the use of traditional method (instead of CNN), and also stating that I did not mention a few technical points in another question. I replied to this email explaining that the choice of traditional method was made owing to its better performance and showing where mention the points they pointed out. In response they said ' the images don't have the same resolution. (means I reduced the resolution before processing the images) Also, the results obtained with cross validation cannot be used to measure generalisation performance.' Could you imagine this coming from someone working in computer vision industry??? This was the biggest turnoff for me. I wouldn't have regretted spending a week on their test had I received a constructive feedback. But providing wrong feedback just for the sack of rejecting is absolutely unethical. Candidates at least deserve a fair evaluation.