It was a non technical, informal chat with a senior data scientist. Mainly to see if you would be a good fit in the role and to tell you a bit more about the company. Very friendly and seemed like a great place to work.

The interview process was highly engaging and well-structured. I began my live coding exercise promptly at 10:30 and was allocated 25 minutes to identify and document bugs in the code. While my task was primarily to document the fixes, I went a step further by resolving all the bugs and successfully running the code. Although most issues were relatively straightforward, the more challenging aspect involved ensuring that the trained model could make accurate predictions on the test data. This was complicated by a mismatch in columns between the training and testing datasets, which required careful attention to detail, especially under the pressure of the interview. After completing the exercise, we reviewed the decisions I made, which were mostly standard. However, there was one question that initially caught me off guard: "What does 100% model accuracy signify, and why could it be a red flag?" My initial response focused on the concept of overfitting and testing on the training set. However, the interviewer was looking for a more mathematical explanation, which I was able to provide with some guidance. Although I hadn't slept for over 24 hours before the interview, which may have impacted my ability to fully articulate my reasoning, I was pleased with my overall performance. Feedback: It was a positive and fair interview experience. My only recommendation would be to include at least two members on the technical panel during the live task or conduct more than one technical interview to ensure a more comprehensive and unbiased assessment of the candidate(reducing the chances of false_negatives from the interviewer's side).