Yappn
Year
2025
Category
AI Web App Design
Role
Product & Visual Designer
Duration
8 Weeks
Problem
The Spark
After spending 48 intense hours completing a take-home coding challenge and acing the first three technical interviews, my husband didn’t get the job. He failed the final round — the behavioral interview.
It felt unfair. His programming skills were undeniable, and his coworkers often praised his reliability and strong collaboration at work. But when faced with strangers asking open-ended questions, his confidence faded. His answers didn’t reflect who he truly was.
Watching this unfold made us wonder:
How many talented people miss opportunities, not because of ability, but because of how they communicate under pressure?
Looking for Solutions
Platforms like ADPList connect people with mentors for mock interviews. It helped — but finding time that fit both schedules was tricky, and even then, talking to a stranger felt intimidating. The problem wasn’t a lack of resources — it was comfort and confidence in the moment.
A Thought
What if you could practice conversations safely, with an AI persona that gives real-time feedback on your tone, clarity, and confidence — preparing you before the real thing?
A space where you can rehearse, reflect, and grow — one chat at a time.
Process
Make strategy -> this is only a proof of concept project, need to ship fast, validate idea through a small group of people(to gain feedback & control the cost)
Potential use cases
Features -> Narrow scope
“Neque porro quisquam est, qui dolorem ipsum quia dolor sit amet, consectetur, adipisci velit, sed quia non numquam eius modi tempora incidunt ut labore et dolore magnam aliquam quaerat voluptatem.”
Whiteboard brainstorming-> High-level structure -> user flow
Use a Gantt chart for the timeline/task management
Highlights
Start to design from the main features and iterate at the same time(usually will be ahead of the engineer, but also need to check with him all the time, cuz we might think differently about a certain feature.)
Dogfooding (noticed some bugs and user experience problems, but we only have limited time, my engineer thinks it's not necessary to make the response faster, smoother, users need to wait a couple of seconds now, I want it to be instant). The LLM audio api he was using is a little slow. He needs to call open ai api, and also needs to think about the cost. At this point, we didn't change anything; we decided to wait for the potential user's feedback.
User testing (gain feedback: how to make the user experience smoother? e.g., the response time is raising the user's frustration level due to the facial expression observation, so we need to improve this. Another problem is that the user didn't seem to notice there's a whole panel that can provide him with real-time feedback; we need to make it more obvious. The purpose of the real-time feedback is that we want the user to be able to adjust their tone, confidence, and fluency cadence along the way, to form better communication habits.
Before:
After:
Iterate the design
Another user testing is smoother.
Design
Other visual deliveries:
Challenge
Challenges we faced and how we solved them
What did we learn?
What did we do better than last time? Last time, from the engineer’s perspective, he was a full-stack engineer who was more passionate about back-end and system design. So, when building the front end, he didn’t have a clear structure. For example, we didn’t focus on reusable/global components or colors from the start because we prioritized speed. As a result, I didn’t build a design library initially. This time, we spent time on components, which saved a lot of time later (colors/font sizes -> tailwind). So, when building our next product, which uses Swift, since we already established our SOP, collaboration will be much smoother.
As a designer,
Product management/timeline management.
