Many students practice for interviews alone and are not sure if their answers are good or not. I wanted a simple tool that gives quick, honest feedback instead of just question lists.

Interview Prep Assistant lets you enter a target company and role, then generates realistic interview questions for that combination. You type your answer, and the app uses Google Gemini to score it on confidence, technical accuracy, and communication. It also points out a few strengths, areas to improve, and suggests a clearer version of the answer.

I built the project during the hackathon using a React frontend and a Node.js backend that talks to the Gemini API. The main challenge was getting stable JSON responses from the model and making the UI fast enough so it still feels responsive while the analysis runs. I learned how to design prompts for structured output and how to connect a full‑stack app to an AI model safely using environment variables for the API key.

Built With

Share this project:

Updates