Hi all,

This is not an SDE role but I used leetcode to prepare. I hope it is okay to post.

Position : New grad in Data and Applied Sciences
Location : Seattle, US.

Stage 1 (Nov 9) : Phone screen (30 min)
Interviewer was a data scientist. They asked about Linear Regression, L1/L2 regularization, over/under fitting, a bit on resume.

Stage 2 (Jan 13) : Final interview (3 interviewes of 45 min each)

Interview 1 focused on

• prediction methods using geospatial data, (answered using locally weighted linear regression)
• fibonacci number (recursive + iterative, with and without memoisation)

Interview 2 focused on

• k means implementation + followup questions on convergence, hyper-parameter tuning
• how do you deal with class imbalnces?
• Bagging and boosting

Interview 3 focused on

• Solve a multi-label classification problem
• Lots of follow-up questions using activation functions,
• Attention mechanism in transformers,
• when will you use custom encoder vs BERT?
• difference between word2vec and BERT?
• difference between self-attention vs relative attention in transformers
• course schedule 2 in last 8 minutes. (did it)

Stage 3 (Jan 25) : Follow up intervew (45 min but went to 75 min)
Focused on

• design a book recommendation system (asnwered using Graph Neural network)
• explain a project + lots of followup questions (trying to apply this method to other products)
• have you worked on a project where you had to change your requirements mid-way?
• how will you make sure that deadlines are met? (answered as a employee but he wanted the answer as a leader)
• do you want to create your own startup someday?
• tell about a failed project.

Received automated rejection on Jan 26.
Thank you for your comments and feedback. Wish you all good luck for your interviews.