Your question is Design Online and Batch ML Serving. 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).
You are building an AI voice platform with personalization across discovery, voice selection, and content recommendations. Some predictions must react to fresh user behavior, while others can be precomputed and served cheaply.
How do you design online versus batch serving for an AI product?