The interview process was quite straightforward and likely the easiest I have encountered. Unfortunately, I didn’t perform well due to a brain fade and neglecting to review my fundamentals beforehand. I concentrated too much on TensorFlow and transformer architecture, which were not part of the evaluation. 1st Round: 30-minute call with the hiring manager. The discussion focused on the role and my experience with Git. 2nd Round: Machine Learning case review/task. The case study was based on the Titanic dataset. Tasks included identifying mistakes in the code and analyzing the train and test datasets. Time allocated: 25 minutes for the task and 25 minutes to review the findings. Code Review: Use of one-hot encoding instead of ordinal encoding. Correcting calculations for accuracy, precision, and recall, and explaining them. Ensuring the target variable is dropped before training the model. Training the model on the training dataset instead of the test dataset. Justifying the choice of the model (random forest classifier) and the number of estimators (trees). Using a validation dataset before predicting on the test dataset. Addressing label mismatches between the train and test datasets (additional column in the train dataset). Handling missing values in the 'age' feature. As mentioned, it was very straightforward. I recommend brushing up on scikit-learn machine learning packages and their hyperparameters, as well as gaining a solid overview of various models such as GLMs, Trees, and SVMs. 3rd Round: Would have been an in-person interview with the partners.