IIEPA: Intelligent Interview & Exam Prep Agent

interface or home page
adding the resume for targeting the skills
adding the target job description for precise targeting
the output analysis
the output and matplotlib donut chart

Inspiration

As a final-year engineering student and UPSC aspirant, I realized that preparing for different roles (like Software Engineering vs. Customer Support) requires completely different strategies. Students drown in generic advice. I wanted to build an agent that treats my specific resume and my target syllabus as the ground truth, offering a hyper-personalized mentor that is available 24/7.

What it does

I wanted to build an agent that treats my specific resume and my target syllabus as the ground truth, offering a hyper-personalized mentor that is available 24/7. (Intelligent Interview & Exam Prep Agent) is an intelligent agent that streamlines the entire preparation process: Ingestion: It analyzes uploaded documents (Resumes, Job Descriptions, or Exam Syllabuses). Strategy: It generates a custom study schedule (e.g., a 90-day roadmap for UPSC or a 1-week sprint for a coding interview). Simulation: It conducts real-time mock interviews with voice/text feedback, tailored to the specific role. Feedback: It provides instant grading on answers, highlighting weak areas and suggesting improvements.

How we built it

Backend: Python AI Engine: Google Gemini API (used for reasoning, question generation, and evaluating user answers). Interface: [e.g., Streamlit / Flask / React] for the user dashboard. Data Processing: [e.g., PyPDF2] for parsing PDF resumes and syllabuses.

Challenges we ran into

One major challenge was ensuring the AI didn't just ask generic questions. We had to prompt-engineer the Gemini model to strictly strictly adhere to the uploaded context (the resume or syllabus) to ensure the mock interview felt realistic and difficult.

Accomplishments that we're proud of

We successfully integrated the "Mock Interview" loop where the AI remembers previous context , it feels like a real conversation, not just a static quiz.