I was contacted by a recruiter.

In the first round, the interviewer asked in detail about my projects, my interests, and why I was interested in joining Morgan Stanley. The interviewer told me they would be hiring me for a machine learning position and explained in detail which machine learning problems their team was trying to solve.

Round 2 was onsite at their New York Office.

I was told I would be interviewed for a machine learning role, but the interview was totally unorganized. There were a total of 3 rounds. None of the rounds had any structure, and not even a single ML-related problem or any algorithms and data structure questions were asked.

Most of the questions were from operating systems, databases, and Java. It was a very bad experience; I still don't understand on what criteria they were assessing the candidate. I did answer the questions but received a rejection two weeks later.

Various Linux commands and operating system questions, such as: What is the kernel? What are processes and threads?

A couple of questions on SQL joins and what the result would be for each join operation.

Java object-oriented principles: What are abstraction and inheritance? What is the use of the super keyword, interfaces, etc.? Design a class using all OOP principles and explain why you chose to design it this way.

Compute the square root of a number without using mathematical operators.

In a CSV, certain columns are stored. Read the CSV, perform certain join operations, and return the result.