I applied online. The process took 2 weeks. Initial phone screening, coding screening, problem based technical interview, then a conversation with senior management. Coding portion isn't too hard, but you will have to have an okay conceptual grasp on traditional ML models and more modern NN approaches. Technical interview was problem based. They introduced a problem, you have to provide a high level solution. They asked some model design questions, stuff about data validation, performance monitoring, etc.

How would you improve this model implementation? (i.e. to improve scalability, model performance, supporting different/more complex inputs)