Your question is Real-Time Multimodal Telemetry Analysis. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
Design a system to process and analyze multi-modal telemetry data (logs, metrics, and traces) in real-time at Datadog.
Clarify the detection and analysis objectives, then propose an end-to-end ML system covering ingestion, feature generation, anomaly detection, cross-signal correlation, serving, and feedback. Discuss scale assumptions, latency, freshness, evaluation, feature drift, training-serving skew, and failure recovery.