I would not recommend to apply unless they change their hiring procedure. As it is at the moment, they are using the hiring procedure to get unpaid work and ideas from the applicants. Stage 1, before getting the chance to talk to anyone and to ask questions about the company: read 6 papers (10 to ~50 pages), find a recent paper relating to one of the topics, write 1-2 pages including a summary of the paper, how it relates to the literature, what it adds, pros and cons and how it relates to the company's products. Note: if they don't like your solution you won't get any kind of feedback, just a standard rejection mail. Stage 2 & 3: both similar, present your results to one team member each followed by a couple of questions relating to the paper you chose and how the company can make use of the results. In my case, we barely talked about my background or prior experience. First interview was good, the person I talked to in the second was extremely condescending and clearly had no idea how to conduct an interview. They did not get back to me within the agreed time frame and I had to chase them down. I got a rejection based on a lack of experience in a specific area. It is more than obvious from my CV that I do not have that particular experience. It does not seem like they ever seriously considered hiring me. They absolutely minimise the time they spend on one applicant while asking you to put in several hours of work before even getting the chance to talk to someone and ask questions about the company or position. Even if they already know based on your CV that they do not want to hire you, you will still be asked to do the research task and present your ideas for their products. Hence, I would not recommend anyone to waste their time with an application to neurocat.

2 options were possible for the technical interview: - Summarize and discuss a research article on AI robustness or explainability; - Implement basic adversarial attacks on a pre-trained open-source object detection model. Then there were informal interviews with the research lead and with HR.