What are your strengths and weaknesses?
Associate Scientist I Interview Questions
820 associate scientist i interview questions shared by candidates
Machine Learning Theory Interview (By Hiring Manager) 1. Interviewer started asking about my thesis and asked questions about IMU sensor, Odometry 2. Classification metrics (Precision, Recall, curves - what is measured along X, Y and how is area calculated, how you conclude about best threshold) 3. Decision Trees and Random forest regularization techniques 4. Advantages of LLM over LSTM 5. Explain LLM architecture and every layer functionalities 6. Scenario-based question: How would you ensure a particular class is distinguished correctly during training and achieves zero errors? (By assigning a higher weight to that class during training.) 7. Explain about Adaboost and how you can modify according to your dataset 8. Scenario based: If the Product Manager provides a new requirement, how would you approach and execute it? 9. Scenario based: Explain a scenario from your prev experience where you had to work with multiple stakeholders 10. Scenario based: You are given a requirement to develop a solution that relies on external AI APIs. However, the Product Manager does not want any sensitive data to be sent outside the system. How would you handle this situation? I would initially use the external APIs but ensure that all sensitive user data is masked or anonymized before sending any requests. This allows rapid prototyping while keeping data privacy intact. In parallel, I would begin working on developing an in-house model capable of handling the required tasks, so that we can eventually migrate away from external APIs entirely. (This was the same approach eventually adopted by the team.)
Top Grading Discussion (By Director - Applied Science) This round involved a deep dive into my projects, my role at Walmart, and my hackathon experiences. The interviewer focused on understanding my overall approach, problem-solving process, and decision-making rationale. I was asked questions such as why I selected particular models, what training strategies were used for fine-tuning, how knowledge distillation was applied, and other aspects that revealed technical depth and reasoning.
Why did you choose this model How did you process missing values
some questions related to sth which are confidential
biggest weakness (as usual)
What new skills would I bring to the group?
Talk about the XGbooster. And how to use it to tuning hyper-parameters.
Tell me about yourself and your experiences
How would you verify the system suitability of an LCMS run?
Viewing 781 - 790 interview questions