BBYou are designing the ML system for a real-time chat application. The product needs to surface the most relevant conversations and message suggestions while keeping responses fast and fresh as users send messages continuously.
How would you implement a real-time chat application?
