This was one of the worst interview experiences I have ever had. I was initially contacted by a recruiter from Talentful, a company that Expedia apparently uses for hiring, who scheduled a first interview with the Hiring Manager.

It was supposed to be an "informal chat," as the recruiter mentioned, but it was actually a proper technical interview. Nevertheless, I was moved to the next stage.

The next stage was supposed to be a take-home assignment about hotel ranking (Kaggle competition style), but the recruiter failed to send the link to access the project. At the very last minute, this step was replaced by a live coding session during the onsite interview (the next step).

Finally, I completed the onsite, which consisted of four one-hour interview sessions: live coding, statistics, machine learning theory, and machine learning system design.

The live coding session included two exercises (one about linked lists and the other about sorting/hashing). The statistics interview was more like a "traditional ML" interview, with some focus on hypothesis testing and distributions. The machine learning theory was actually very much applied to a real ranking problem, and the final ML system design was more about general production and architecture settings given a problem.

Again, the recruiter failed to give any advice on any of the interviews, so I had to prepare completely blind, relying only on my past interview experience. I have had interviews with other big tech companies, and all of them provide a much richer set of information to prepare for the interviews.

At the end, I still hoped to get the job because I had very nice chats with the people interviewing me, who all looked skilled and professional. I apparently did okay but was rejected due to a lack of experience (something that could have been seen from my CV during the screening, I guess...).

I still consider Expedia a nice place to work, but they really need to do something about the recruiters' incompetence.

The whole process lasted about two months, with a lack of communication and very poor timing from the recruiter's side. Better communication from the recruitment team is strongly advised.

Talk about the bias-variance trade-off.

How would you build a recommender system for the Expedia webpage?

Talk about p-value, hypothesis testing, and A/B testing.

What metrics would you use for a ranking algorithm?

General deep learning questions (dropout, regularization, overfitting).

When would you choose a Random Forest or a Logistic Regression model?

Advantages of Logistic Regression versus Random Forest.