The first stage was an OA that took about an hour. It had two parts. The first part was maybe a dozen or so multiple choice questions, covering a lot of ground quickly — there were data processing questions, questions about statistical distributions, and some basic hypothesis testing concepts. There was plenty of time, I think it was 25 minutes for under 15 questions?
The second part was rough — it was a modeling task. The question was a CSV file about red wine, where the dependent variable was wine quality (something like a 1-7 scale), and there were around 10 independent variables. There were about four sub-questions, and the first two were EDA questions — like what did you find in the data, which variables could be used, and so on. The data was clean; the prompt said upfront there was no missing data or anything like that.
You could pick your own model. The goal was to predict wine quality and figure out which variables mattered. The problem was that when I got the OA, nothing said I'd need to build a model, and the platform didn't give you an IDE either, so I panicked and spent about 10 minutes setting up a notebook locally — 40 minutes total. The model I ended up building was garbage, accuracy was only around 60%, and in the end I didn't have time to upload the notebook.
The surprising part is that I still got an HM interview afterward, all behavioral questions. I'm guessing it still wasn't a good fit, and I got rejected.
Discussion
Loading comments…