Your question is Classify Meeting Feedback by Intent. 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 NLP workflow for a collaboration platform that receives thousands of free-text comments each day from post-meeting surveys, support chats, and feature request forms. Product and operations teams want a single model that can route each message into intents such as bug report, feature request, billing issue, usability complaint, or general praise so they can prioritize follow-up work. You have about 300,000 historical messages with noisy human-applied tags, many short texts with typos and product-specific terms, and a smaller set of long-form comments that reference meeting recordings, chat, phone, and webinar workflows. The system needs to be easy to retrain as taxonomy definitions evolve.
Describe an NLP project you have worked on that is similar in spirit, and explain how you would design and implement this intent classification system end to end. Walk through the preprocessing pipeline, model choice, training approach, and how you would evaluate and improve it in production.