ML questions were very easy, like PCA, l1 and l2 regularization, roc code for svm, p value, basic scenario based questions on imbalance classification. They said there will be discussion round for those questions but it didn't happen. A case study was based on simple profit loss, and if you do one simple mathematical mistake due to nervousness, they will say to you to leave.
sql codes for different cases
LeetCode-style question of easy/medium level.
Explain how you would handle missing data in a dataset for a machine learning model.
tell me more about yourself
1) Coding questions Leetcode Difficulty level Medium 2) Technical background, work experience
Explain what is embeddings and how to convert to vectors
How would you define MLOps?
HR screen was simple asking me to walk through my resume. The coding challenge was difficult but, not too much. Ford does not allow any coding language besides Python. However, the grading of the coding challenge is a problem area. I answered all 3 questions and provided the data plots and accuracy metrics in the code. But to my dismay, I discovered later that they graded me a 7% score. This was beyond belief. The recruiter refused to even cross-check if there was any s/w glitch or, human error in the exchange of files or, if there could be any problem with the HackerRank s/w itself. I was declared to be a duffer and written off. I would never interview with Ford again.
Tell me a moment you achieved sth very important for your company
Viewing 1611 - 1620 interview questions