I applied online. The process took 6 weeks. I interviewed at SAP (Walldorf, Baden-Wurttemberg) in Sep 2017
Interview
* fast reply within 2 weeks after sending the application
* very nice and professional contact with HR
* invited to short screening phone interview (~15min., "introduce yourself", got information about the position and application process)
* offline ML challenge, 1 week to solve, interesting and fun problem
* afterwards invite to 1 hour phone interview with 2 experts of the team, very nice guys, asked in detail about my previous ML experiences, gave a typical ML problem and asked how I would approach it ("what data would you ask for", "what model would you build", etc.)
* detailed questions about deep learning, e.g. properties of different activation functions, advantages/disadvantages of several methods etc.
* at the end we had a relaxed conversation about the team and the job
* 2 weeks after HR let me know that they are not considering me anymore
Interview questions [1]
Question 1
Why did you choose LSTMs instead of more basic RNN architectures?
I applied through a recruiter. I interviewed at SAP (Prague)
Interview
In general the interview process was smooth and the recruiter Sylvie was very good and informative on all the steps of this process. I wouldn't say the same about the people that did the technical interview as it was more difficult to communicate with them and have a proper conversation.
Dynamic programming question to be solved on a text editor explaining the logical reasoning for the answer. Afterwards further clarification questions. Also explanation of time and memory complexity. That is to explain the Big O complexity for time and memory.
Interview questions [1]
Question 1
Dynamic programming question to be solved on a text editor explaining the logical reasoning for the answer.
1st round- online technical assessment wasn't too bad (python, timed, focused on linear regression)
2nd round-behavioral (thought i did well, but didn't make it past)
3rd round- technical panel with whiteboard data structure type problems, situational
