Can't recall the specific questions and they varied by individual. The worst questions were from someone who grilled me on steps of analyses and actually said "pretend you are walking to your computer. Now tell me how you are going to analyze your data." Others were more innocuous but there was still an element of testing ad nauseum.
What project did you deliver so far?
Explain Backpropagation and it's advantages.
Given that we have a machine learning model that performs well locally but when deployed to users doesn't what could possible be the cause of this
alot on resume and theory on concepts covered in resume
🔹 1. Conceptual Questions (Beginner–Intermediate) ❓ Supervised vs. Unsupervised learning What is the difference between supervised and unsupervised learning? Give examples of real-world problems for each. ❓ Model Understanding What is overfitting and underfitting? How do you prevent overfitting? What is the bias-variance trade-off? What are precision, recall, F1-score, and when do you prefer one over another? ❓ Algorithms How does a decision tree work? What is the difference between logistic regression and linear regression? How does K-nearest neighbors (KNN) work? What is regularization (L1 vs. L2)? 🔹 2. Intermediate to Advanced Topics ❓ Ensemble Methods How does random forest work? What is gradient boosting (e.g., XGBoost, LightGBM)? Difference between bagging and boosting? ❓ Neural Networks What is backpropagation? What are activation functions and why are they important? Difference between CNNs and RNNs. What is dropout, and why is it used? ❓ Optimization What are common optimizers in deep learning? How does stochastic gradient descent (SGD) differ from batch gradient descent?
Why do you want to join Microsoft?
Self introduction and my work experience.
Coding test task 1: pandas on what looked like a phone company database (super lengthy) Coding test task 2: lasso, ridge and elastic net Screening: questions about past experience
Explain what the following code segment does, and when someone would use this segment in a real application (included a lengthy discussion about performance metrics)
Viewing 1141 - 1150 interview questions