I applied through a recruiter. The process took 3 weeks. I interviewed at Block in Apr 2021
Interview
The recruiters do their job very very poorly.
I got contacted by the Square recruiter through LinkedIn:
Initially, the recruiter asked which role I want to proceed (ML modeler or Data Scientist). I said I am ok with both but prefer DS. Then, I started its process and passed two other technical interviews afterward (Python coding & SQL, both quite easy). But, the recruiter came and said you have to change to ML modeler as your profile is more suitable for that role (so why starting from DS in the beginning?!). Then, I had another interview with hiring managers both from Growth and Risk (two different streams at Square). However, the recruiter forgot to attach the google meet link. I ended up talking to one of the hiring managers on the phone, which I rather have it on a video call, instead. Then, I passed and went to the final step, virtual onsite. I was more interested to ML modeler in Risk, and I expressed my interest to recruiters, but all my interviews were set with Growth managers, and recruiters keep lying to me that you are interviewing for both streams. For the virtual onsite, I was booked for four interviews in the invitation description (for the first day, followed by another three interviews for the next day). However, I ended up having an additional interview on the first day that I didn't expect at all, which was actually the most important part.
In every single stage, recruiters were very disappointing. They fail to do their only job!
Very lengthy process:
recruiter initial call
python pair programming
SQL programming
hiring managers (two interviews)
virtual onsite day 1 (5 interviews, 4 hrs)
virtual onsite day 2 (3 interviews, 2 hrs)
Most of the interviewers were cool and friendly, they help you in the process with tips and hints. Questions were fairly easy to solve. Hiring managers did not connect with people. They don't see interviews as a two-sided evaluation process.
Interview questions [1]
Question 1
lots of case studies, hypothesis testing, pair programming for fraud classification