Mockr

Inspiration

With thousands of students competing for the same internships, one of the biggest differentiating factors is how well you perform in the interview. We decided to make Mockr to help users truly stand out when they get to the interview phase.

What it does

Our app is a mock interview simulator that allows users to practice both technical and behavioural problems. The web app has an integrated sandbox for users to code out their solutions and the AI will provide feedback after the interview is complete.

How we built it

We used typescript for the front end, opencv for head detection to track non verbal cues, elevenlabs for the AI voices, and openrouter for our AI LLM that responds that acts as the interviewer.

Challenges we ran into

A big challenge we ran into was the limited access we got to the different API's we needed. Originally, we used Gemini API and Elevenlabs API but they both ran out of tokens quite quickly. In order to solve this, we switched to really low cost platforms, using the free credits we got for openrouter and switching to a local text to voice model in order to let us freely test the web app.

Accomplishments that we're proud of

We're proud of completing this project in under 24 hours. It was quite an impressive achievement to do it under such time constraints.

What we learned

We learned that some of the seemingly easiest tasks ended up being the most problematic to be fixed. For example, we thought that implementing the elevenlabs API would be a simple call to the API in the front end but then we realized we needed to set up an entire backend to handle the calls for the elevenlabs API to work.

What's next for Mockr

We hope to continue expanding the types of interview questions as well as hopefully adding an AI interviewer that can video call with you to more closely simulate an interview experience.