First Round- Basic Aptitude, some mental ability questions and basic HR questions.
Second Round- Machine learning and ML system design interview. Questions would mostly come from NLP like explaining word2vec, and what semantics means in NLP. Then some questions would be based on the Resume(better stick to simpler projects, as my interviewer, was a bit impatient and my highlighted project was a bit complex). Next, it will be followed by ML systems design questions like what can kind of analytics and insights, you can get from chatbot data of a product company. After that there will be a rapid-fire round with random questions from statistics, basic ML like what is top-down and down-top unsupervised learning techniques, given a matrix what is its rank, difference between vector and tensor (bit incoherent set of questions would be thrown on you)
3rd Round Programming Aptitude round. This round will start with an easy-medium programming question with different variations of it. For example, given a sorted array and a threshold find the elements to the right of the threshold and left of the threshold (keep in mind that there may be duplicate elements in the array). Followed by some OS-related questions like concurrency and parallelism. Thread and process. Then whether threads are processed parallelly or concurrently. Then some basic questions like how does garbage collector in python work and what is its disadvantage. Followed by a system design question about designing a database for a restaurant ordering system (taking care of things like priority, batch orders, optimization by clubbing food orders etc).
There may be more rounds after this I don't know about. I had these 3 rounds then I was rejected. This interview was for an entry-level role (MLE 1)