NopiroAI

User login
User input
nopiro.vip website (pre-launch)
Interface before countdown to interview
Interview interface

Inspiration

This project was inspired by our own experiences during the college application process. We realized that many high schoolers lacked any interview experience during their high school years, leaving them ill-prepared for important scenarios that can significantly impact their lives. Not every student has access to real interview training — such opportunities can be costly or difficult to find. We wanted to develop a resource that allowed students to practice important questions to help them feel more confident and prepared.

What it does

Nopiro uses user inputs (university name/major) to personalize a college application interview similar to what one might experience in real life. An AI avatar will interview the student using a streaming API to increase the authenticity of the experience (unlike some interview prep resources, which require the user to record a response, Nopiro converses with the user in real time). This real-time interaction simulates the spontaneity of a real interviewer, exposing users to unexpected follow-up questions and helping them practice more natural and fluent responses. It delivers an immersive and realistic interview experience—while keeping the psychological pressure lower than in a live, human-conducted interview.

How we built it

We built Nopiro using Next.js for the core web framework, Tailwind CSS for responsive UI design, and Vercel for hosting and deployment. The platform integrates Snowflake for scalable data storage and retrieval, Gemini API for dynamic AI‑driven interview questions and responses, and Eleven Labs API for generating realistic voice output to power the avatar’s speech and enhance user interaction.

Challenges we ran into

During this hackathon, we experienced no shortage of challenges; as a group of first-time hackers, we were navigating the process of creating a team project while learning and implementing different technologies. While creating small aspects of Nopiro (Using api keys, coding the UI, etc.) was simple, finding a way to connect everything while not leaving any holes took a lot of communication and time (mentors were a great help during this process).

Accomplishments that we're proud of

As a team, we are proud of the fact that we were able to complete the entire process (idea, drafting, building, honing) and produce a final product that was what we envisioned.

What we learned

Most of the processes that we used to build Nopiro were new for us, from combining STTs and TTSs, to avatars and lip syncing, to specific things like streaming APIs; we gained an abundance of skills during the development of Nopiro.

What's next for Nopiro

Due to its scalability, Nopiro has the potential to be used as a mock interview tool for other kinds of interviews (technical, behavioral, etc.) Its use of Snowflake to store data and its versatility using an AI avatar and Gemini can allow it to operate as any personalized communicator (tutor, Q&A, assessments, etc.) as long as it has data. As the user base grows, Nopiro can continuously collect feedback and performance data to further improve its conversational quality. This allows the system to generate questions and feedback that are increasingly tailored to each user’s unique background and progress, making every session more relevant and effective.