Recruiting for a software role? Building a product? Have a repetitive workflow that should probably be code? I’m interested.
Or, if the conversation is about Formula One, you already have my attention.
Abhinav Mishra
BACKEND-LEANING SOFTWARE ENGINEER
PrepEdge AI is a full-stack, AI-powered mock interview platform designed to make interview preparation more realistic, personalized, and measurable. Instead of practicing from a generic list of questions, users can create mock interviews based on their resume, target role, job description, and interview type. The platform generates relevant questions, evaluates responses using AI, provides detailed feedback and scoring, and helps users understand where they need to improve.
Traditional interview preparation often revolves around reading question lists or practicing answers without receiving meaningful feedback. This makes it difficult to know whether an answer is actually good, whether important points were missed, or how prepared someone really is for a particular role. PrepEdge AI attempts to close this feedback loop by creating role-specific interview experiences and evaluating responses immediately. It helps candidates practice in an environment that is closer to an actual interview while continuously identifying weak areas and tracking improvement.
Users can expect personalized AI-generated interview questions derived from their resume, role, job description, and selected interview type. Responses can be evaluated with per-question scores and detailed AI feedback, followed by an overall interview summary. The platform also supports voice input and speech analysis, reusable interview templates, performance analytics showing score trends and weak topics, and downloadable PDF reports that can optionally be shared through public links. The goal is not simply to ask questions, but to create a complete practice-feedback-improvement cycle around interview preparation.
Frontend: React 19, Vite 6, Tailwind CSS v4, TanStack Query and React Router 7.
Backend: Node.js with Express 5, Mongoose and Firebase Admin.
Database: MongoDB Atlas.
Authentication: Firebase Authentication.
AI: Groq, Google Gemini and Hugging Face using task-specific fallback chains, with Groq Whisper used for speech-to-text and voice-related functionality.
Deployment: Vercel for the frontend and Render for the backend API.
The idea for PrepEdge AI came from my own interview experience. My first interview did not go the way I expected. I was nervous and, more importantly, I had very little idea about the kind of questions I could actually be asked or how well my answers would perform in a real interview.
While preparing afterwards, I realized that reading interview questions was not enough — what I really wanted was a way to simulate the experience, receive feedback, and understand exactly where I was going wrong.
That experience became the motivation behind PrepEdge AI: if interview preparation could be personalized around a candidate's own resume and the job they were targeting, while also providing instant feedback, candidates could walk into real interviews significantly better prepared and more confident.
PrepEdge AI grew from a personal project into an open-source project selected for GirlScript Summer of Code (GSSoC) 2025, where developers were able to contribute to and improve the platform.
As of September 2026, the public GitHub repository has earned 35 stars, 36 forks and accumulated more than 160 commits. The repository also has an active issues and pull-request history, showing that the project has attracted participation beyond its original creator.
Its inclusion in GSSoC and the community activity around the repository are especially meaningful because the project originally started as an attempt to solve a problem I personally experienced during interview preparation.