Data Scientist / ML Engineer interviews are divided into 3-4 rounds (5 in case of FAANG). Here are the details about each round:
1. ML theoretical depth and breadth:
In this round, you will be tested with pure theoretical questions from statistics, math, ML, DL, NLP and Gen AI. Also deep dive into projects mentioned in tye CV – interviewer can go in depth for the topic used in project.
2. DSA / ML coding:
In this round, pure DSA questions will be asked ranging from easy to medium difficulty. Then there will be few ML coding – data handling with Pandas, implement forward prop of a NN, write basic RAG architecture etc.
3. ML system design with real business use cases:
This is one of the craziest round. You will be given a real business use case problem and asked to design a solution. Intuition about each and every ML algorithm, evaluation metric, API concept, and deployment knowledge are required.
4. managerial / leadership + behavioral:
This round is less technical but requires knowledge about past projects, how you handled different situations, how to lead a team, and your cultural fit all are required. This round is often taken lightly but many rejections come along this round. Take it seriously.
