README.md

metadata

title: Excel Mock Interviewer
sdk: streamlit
emoji: 🚀
colorFrom: purple
colorTo: purple
sdk_version: 1.49.1

Excel Mock Interviewer (Streamlit PoC)

This repository implements a single-process Streamlit Proof-of-Concept for the AI-Powered Excel Mock Interviewer.

Quickstart (local)

1. Install dependencies

python -m pip install -r requirements.txt

2. Run Streamlit app

streamlit run app.py

3. Open http://localhost:8501

What each file does (step-by-step)

• app.py — Streamlit UI and control; displays questions, captures answers and uploads, shows summary.
• interview_logic.py — Manages interview state, history, session id, and orchestrates evaluation calls.
• evaluator.py — Deterministic autograder helpers using openpyxl and an LLM wrapper (OpenAI when enabled). Returns parsed grading JSON.
• prompts.py — Questions and prompt templates used by the grader.
• requirements.txt — Python deps.

How grading works in the Streamlit PoC

• For text answers: interview_logic calls evaluator.call_llm() which either calls OpenAI (chat completions) to obtain a JSON grade or uses a simple heuristic fallback.

Extending to production

• Replace Gemini(free) grader with a high-quality production LLM and stronger prompt engineering.
• Persist sessions to a database and files to object storage.
• Add authentication and consent screens.
• Implement active-learning pipeline where low-confidence grades are routed to human reviewers.