JOBREADY AI

JOBREADY AI

Demo Video- https://drive.google.com/file/d/14ymS1Qpe5EzWkTfqXM07w7ggT1JL50Nr/view?usp=sharing

Inspiration

As recent graduates actively preparing for full-time opportunities, we realized that interview preparation can be stressful and inconsistent. We were getting only a few real interviews and wanted a way to simulate the real experience with voice, timing, and instant feedback, so that when the actual call comes, we feel confident and ready. This also provides ATS-friendly feedback. That’s how JOBREADY AI was born.

What it does

Generates personalized, role-specific questions (Technical, HR, Managerial, etc.) Provides voice-based AI feedback and scores after each round Uses a virtual interviewer voice (Clara from ElevenLabs) and camera view to simulate real-world pressure Helps users build confidence, refine communication, and identify weak areas before facing recruiters In short, it’s like having your personal AI interviewer available 24/7. Also helps to provide ATS-friendly feedback.

How we built it

Frontend: Built using Streamlit for rapid prototyping and an interactive UI. Voice Engine: Integrated ElevenLabs TTS (Clara voice, eleven_multilingual_v2) for realistic conversational feedback. Backend: Python-based APIs for question generation, voice synthesis, and feedback. LLM: Powered by Gemini(gemini-flash) for generating questions, evaluating answers, and providing insights. Speech Recognition: Used Google Web Speech API for transcribing recorded responses. Resume Parsing: Implemented PyMuPDF, pytesseract, OCR and python-docx to extract data from PDF/DOCX resumes.

Challenges we ran into

Managing voice latency and ensuring natural flow during question-answer transitions. Parsing scanned resumes accurately while preserving formatting. Maintaining conversational flow while switching between text, audio, and feedback modes. Integrating multiple APIs (Gemini, ElevenLabs, Google Speech) seamlessly into one pipeline.

Accomplishments that we're proud of

Created a fully functional AI interviewer that feels natural and engaging. Combined voice, vision, and text into a single interactive platform. Designed a dark-themed, modern UI with custom CSS for an immersive experience. Successfully handled both typed and voice-based answers. Delivered accurate and constructive AI-driven feedback and scoring per question. Built everything from scratch within a limited time and resources, while learning fast!

What we learned

How to design a multi-modal AI system integrating text, voice, and UI interaction. How to handle real-time voice synthesis and speech recognition in a smooth loop. The importance of prompt design for generating accurate, context-specific interview questions. How to make the user interface intuitive and less intimidating for interviewees. Effective teamwork, debugging, and rapid prototyping under hackathon pressure!

What's next for JOBREADY AI

Deploy a React + FastAPI version for scalable web use. Add emotion and tone detection from user voice and face using OpenCV and Azure Speech SDK. Integrate LangChain for deeper context handling and consistent feedback. Include session history so users can track their improvement over time. Add leaderboards and gamified badges to make interview prep engaging. Build a mobile app version for on-the-go mock interviews.