loading
All projects
2025AI-powered mock interview platform with speech transcriptionActively maintained
Mockmate
AI-powered mock interview platform built with the MERN stack. Uses Gemini AI for dynamic question generation and Azure AI for real-time speech-to-text.
Mockmate streamlines software engineering interview preparation by simulating realistic interview boards. Using resume parses and role specifications, Gemini AI dynamically customizes a sequence of behavioral and technical questions, adapting difficulty based on candidate response performance. Candidates answer orally, and their responses are converted into text in real time using Azure Cognitive Services (Speech-to-Text). The backend processes transcripts and streams follow-up questions and live status updates to the client via Socket.io.
The architecture is deployed serverlessly using AWS Lambda and API Gateway for scaling and cost efficiency, with a React + Vite frontend. After completing a session, the system generates comprehensive sentiment analytics, feedback logs, and score breakdowns for communication, technical depth, and clarity.
highlights
- Dynamic question generation and difficulty calibration driven by Gemini AI
- Real-time verbal answer transcription integrated with Azure Cognitive Speech services
- Live session orchestration and synchronization utilizing Socket.io events
- Serverless backend architecture deployed on AWS Lambda with API Gateway
- Analytics reporting package delivering feedback logs and scores on multiple performance dimensions
- React
- Vite
- Node.js
- Express
- MongoDB
- Gemini AI
- Azure Cognitive Services
- AWS Lambda
- Socket.io