Instruction: Outline the types of cyber threats you would target, the data sources you would use, how you would preprocess and model this data, and how you would deploy the system.
Context: The question probes the candidate's ability to apply machine learning to the domain of cybersecurity, requiring knowledge of threat landscapes, data handling, and operational deployment.
Official answer available
Preview the opening of the answer, then unlock the full walkthrough.
I would start from concrete security workflows instead of a generic promise to "detect threats with AI." In an enterprise setting, the highest-value use cases are usually things like phishing detection, anomalous authentication, endpoint behavior analysis, insider-risk signals, and alert prioritization for analysts.
That means the ML system...
medium
medium
medium
hard