Students and early-career professionals often feel underprepared for interviews—not due to lack of ability, but because practicing can be intimidating. Professional career coaching can be expensive, mock interviews can feel awkward, and many candidates fear being judged. Interview Pal provides a safe, realistic, and on-demand environment for interview practice anytime, anywhere.
- Users select a role and company, which the backend matches to a stored job description.
- Users upload their resume, which is parsed for key skills, experiences, and projects.
- The AI generates context-aware, role-specific interview questions.
- Users conduct real-time voice interviews with the AI agent.
- After the session, users receive a detailed evaluation and feedback on performance.
-
Frontend Built with React + Redux, providing a responsive, real-time user interface with seamless voice integration.
-
Backend Built with FastAPI, serving as the API layer for agents, resume parsing, and database operations.
-
Resume Analysis Agent Extracts structured information from resumes using NLP techniques: skills, experience, projects, and achievements.
-
Question Preparation Agent Generates interview questions tailored to the user’s profile and the selected job description using context-aware LLM prompts.
-
Interview Agent Conducts live, real-time voice interviews leveraging the OpenAI Realtime SDK. Maintains persistent session context for coherent, natural dialogue. Handles interruptions and background noise gracefully.
-
Model: gpt-3.5 turbo
-
Other Tools & Framework
- LangChain for orchestrating multi-agent workflows
- Supabase for user data storage and authentication
- NLP for context-aware
- Built a fully real-time, voice-driven interview experience
- Developed resume-aware AI question generation
- Completed the project within 24 hours during an in-person hackathon
- Implement AI-driven interview evaluation with actionable feedback
- Track user progress across multiple sessions
- Expand multilingual support and accessibility
We are currently in the midst of beta development, and have not deployed this thing due to API credit contraints. But, here is a video demo! https://youtu.be/NL71paK-edo