PRESSURE - Train for the moments that matter

Inspiration

Not everyone has access to interview coaching or people to run mock interviews with, which creates an uneven playing field. PRESSURE was built to make realistic interview practice accessible to anyone, regardless of location or economic background.

What it does

PRESSURE is a live, voice-based interview simulator that adapts to the role a user is applying for. It conducts realistic interviews, listens to spoken responses, and evaluates how clearly users explain their thinking under pressure. After the interview, it provides a detailed report with strengths, improvement areas, and actionable feedback.

How we built it

We built PRESSURE as a web-based application using real-time voice input and an AI-driven interview engine. Spoken responses are transcribed and processed to drive a live, turn-based interview experience. The system adapts questions based on the role and interaction flow, while storing transcripts and interaction metadata throughout the session. After the interview, the AI analyzes the full conversation to generate structured feedback and scores.

Challenges we ran into

One of our biggest challenges was keeping the experience fast enough to feel like a real conversation. Early on, responses took too long, so we had to rethink how and when we run analysis to keep the interview feeling live. On top of that, getting the interview questions and overall flow to feel natural took several iterations, and we spent time refining the frontend so the experience felt clean and focused rather than overwhelming.

Accomplishments that we're proud of

We’re proud that we built a complete, end-to-end interview experience that actually feels like a real conversation. PRESSURE listens to spoken responses, responds in real time, and asks follow-up questions when clarification is needed, rather than just moving through a script. The system adapts to the role and interview flow, and generates a detailed post-interview report that users can download and keep. We’re also proud that the experience is accessible as it runs entirely in the browser, doesn’t rely on a camera, and gives anyone a fair way to practice explaining their thinking under real interview pressure.

What we learned

We learned that building applied AI isn’t just about model quality, it’s about system design and user experience. Small architectural decisions, like when to run analysis, had a huge impact on whether the interaction felt natural or broken. We also learned that clear constraints and thoughtful prompting are essential for getting consistent, useful behavior from AI in a live setting. Most importantly, we saw how much trust and clarity matter when AI is directly interacting with people in real time.

What's next for PRESSURE - Train for the moments that matter

While PRESSURE is currently focused on interview preparation, the same system can be applied to other high-stakes conversations where clear communication under pressure matters. Future scenarios could include workplace conversations, conflict resolution, or other difficult discussions that benefit from realistic practice and structured feedback.

Built With

• elevenlabs-api
• fastapi
• framer-motion
• lottie
• next.js
• openrouter
• react
• tailwind-css
• typescript
• web-speech-api