Yoodli analyzes speech recordings and provides real-time feedback on metrics like filler words, pacing, clarity, and delivery. Users record practice sessions (ranging from 2-30 minutes) and receive detailed feedback afterward as the AI processes different aspects of their speech.
The analysis happens asynchronously through multiple AI model calls — some return quickly, others take longer. Users should see feedback appear progressively rather than waiting for all analyses to complete.
Your Task:
Design how you might build the recording analysis page, where feedback streams in over time from multiple AI calls. Consider how to display recording playback, organize feedback by category as it arrives, handle varying recording lengths, and allow users to navigate between feedback and recording moments.
Assumptions You May Make:
Submit a system design sketch and video response walking us through your approach.
Demonstrate full stack engineering fundamentals
Design scalable system architecture
Handle asynchronous data flows
Create smooth user experiences with progressive loading
Balance technical tradeoffs in a startup environment
On this page
