Ho presentato la mia candidatura online. Ho sostenuto un colloquio presso Microsoft (Seattle, WA) nel mese di set 2025
Colloquio
The first-round interview was scheduled for 30 minutes but lasted about 45 minutes. It was a phone interview. The interviewer started with a few basic machine learning questions. After that, I was given an algorithm problem to answer verbally without coding. Then I was asked about challenges I faced in my past projects and finally to describe one recent research project that impressed me the most.
Domande di colloquio [1]
Domanda 1
Q: Can you explain the bias-variance tradeoff in machine learning?
Q: [Algorithm question] easy to medium level difficulty
Q: What challenges did you face in your projects?
Q: Can you describe a recent research project that you found memorable?
Ho presentato la mia candidatura online. Ho sostenuto un colloquio presso Microsoft
Colloquio
The interview process started with a short 30-minute introductory interview, where we mainly discussed my previous experience, background, and motivation for the role. After that, I had three interviews in one day. One interview focused more on my previous projects and experience, while the other two were more technical and focused on machine learning theory, practical ML problem-solving, and model selection. The interviewers were very nice, professional, and the atmosphere was comfortable throughout the process.
Domande di colloquio [2]
Domanda 1
Describe how you would build a model that recognizes whether a person is wearing glasses in a photo.
Two rounds of interview, both technical, in both rounds I was asked about my projects and very deep insights related to those, also the questions were very practical based ml questions, i was asked to do live coding also in both the rounds
Ho presentato la mia candidatura online. Ho sostenuto un colloquio presso Microsoft nel mese di mar 2026
Colloquio
3 interviews. One ML breadth mostly focused on Transformers. What are hallucinations, and how to avoid them, what is catastrophic forgetting, and similar questions. Another manager round with some HR questions, like sharing past learning experiences, was also asked about an open-ended ML system design for next-word prediction in an app. Third was the coding round, where they asked me to find an accurate threshold for classifying using a really bad classifier. The answer should be the F1 score, and I was supposed to code it up live.
Domande di colloquio [1]
Domanda 1
What are hallucinations, and how to avoid them, what is catastrophic forgetting, and similar questions. Another manager round with some HR questions, like sharing past learning experiences, was also asked about an open-ended ML system design for next-word prediction in an app. Third was the coding round, where they asked me to find an accurate threshold for classifying using a really bad classifier. The answer should be the F1 score, and I was supposed to code it up live.