An AI-powered, closed-loop interview simulator designed to automate rigorous preparation for software engineering roles.
Instead of relying on generic question banks, Interview Copilot dynamically generates a tailored mock interview experience by cross-referencing live job descriptions with your personal resume. It identifies skill gaps, simulates realistic technical and behavioral grilling, and provides real-time, rubric-based feedback.
- Job Specification Extraction: Scrapes and parses target job postings (via URL or raw text) to extract required tech stacks, responsibilities, and seniority levels using a structured LLM pipeline.
- Resume Gap Analysis: Ingests PDF resumes, extracts text, and runs a comparative analysis against the parsed job description to highlight matched skills, missing core tech, and preparation recommendations.
- Contextual Question Generation: Synthesizes the company's real-world engineering constraints, job requirements, and resume gaps to generate highly specific technical and behavioral questions.
- Voice-Native Interface: Integrates OpenAI's Whisper API for seamless, timed voice responses, simulating the pressure of a real verbal interview.
- Automated Evaluation Engine: Scores responses based on clarity, technical accuracy, and structural frameworks (like the STAR method). Provides a numeric breakdown, identifies strengths/weaknesses, and generates an optimized "Rewritten Answer."
- Progress Analytics: Persists evaluation data locally via SQLite, surfacing weak areas and competency trends over time in a dedicated dashboard.
- Frontend: Streamlit (including
st.audio_inputfor native voice capture) - Intelligence Layer: OpenAI API (
gpt-4o-minifor JSON-guaranteed structured data extraction,whisper-1for audio transcription) - Web Scraping:
BeautifulSoup4,requests - Document Processing:
pypdf - Data Persistence: SQLite3
1. Clone the repository
git clone [https://github.com/adi-2254/interview-copilot.git](https://github.com/adi-2254/interview-copilot.git)
cd interview-copilot