- Take home: 2 days to read a newly released Deep Learning paper.
- On site: explain the paper to the hiring manager.
- Evaluation: (1) Did the candidate understand it? (2) Did the candidate go farther, and look into the cited papers and understand that? (3) Can they explain Deep Learning concepts and talk their way through related problems?
In 5 months of searching for a job, this felt like the most relaxed interview, because it was really a conversation between people in the field. It was also the most demanding and interesting job I encountered in these five months. I got the job, but this happened especially because the manager actually did the work of respecting my experience and talking to me, instead of going for the traditional and "safe" BS.
An opposite experience:
Technical Interviewer asks me whiteboard SQL question. I answer it, only to be corrected on a minor point. Turns out, when I tried it out on my own, I was actually correct. This person was to be my manager.
I did get an offer from them, but this and other red flags made me decline.
Takeaway for managers who actually care about hiring well:
- Don't force your candidate through one off tasks that don't represent their day to day. Is your Data Scientist really just running SQL queries and optimizing computational complexity of their algorithms? Isn't the main point that they can think critically?
- Hasn't their previous experience shown they can code? Their Github? Sure, give them a (quick) take home project if you really want to make sure. In the rare case they faked their way through the whole thing and you really couldn't tell after talking to them for a few hours, you will surely notice in the first few weeks.
- Take home: 2 days to read a newly released Deep Learning paper.
- On site: explain the paper to the hiring manager.
- Evaluation: (1) Did the candidate understand it? (2) Did the candidate go farther, and look into the cited papers and understand that? (3) Can they explain Deep Learning concepts and talk their way through related problems?
In 5 months of searching for a job, this felt like the most relaxed interview, because it was really a conversation between people in the field. It was also the most demanding and interesting job I encountered in these five months. I got the job, but this happened especially because the manager actually did the work of respecting my experience and talking to me, instead of going for the traditional and "safe" BS.
An opposite experience:
Technical Interviewer asks me whiteboard SQL question. I answer it, only to be corrected on a minor point. Turns out, when I tried it out on my own, I was actually correct. This person was to be my manager.
I did get an offer from them, but this and other red flags made me decline.
Takeaway for managers who actually care about hiring well:
- Don't force your candidate through one off tasks that don't represent their day to day. Is your Data Scientist really just running SQL queries and optimizing computational complexity of their algorithms? Isn't the main point that they can think critically?
- Hasn't their previous experience shown they can code? Their Github? Sure, give them a (quick) take home project if you really want to make sure. In the rare case they faked their way through the whole thing and you really couldn't tell after talking to them for a few hours, you will surely notice in the first few weeks.