I had a terrible experience interviewing for a data engineer job. The interviewer was 15 minutes late. His Chinese accent was so thick that I could not hear what he was saying and had to kindly ask him to repeat himself.

We discussed some theory questions about MR, Hadoop, and Spark. Then, we had a SQL test. The SQL engine was not working, so I was writing code blind. The test ended on time, and I didn't have enough time to complete the questions properly, even though all my answers would have worked, just not with fancy SQL.

The interviewer also demonstrated very poor knowledge of SQL. For one question, I gave an answer with a self-join that would have worked, but he wanted a window function and claimed my answer would result in a NULL. He was outright wrong; I checked ChatGPT to confirm my answer, and it would have worked.

Past projects with ETL.

Theory questions about MapReduce, Hadoop, and Spark.

Some of these questions were quite obscure, like:

How does Spark process skewed data?