README.md

---
title: Codingo
emoji: 🤖
colorFrom: indigo
colorTo: blue
sdk: docker
app_file: app.py
pinned: false
---

# Codingo — AI-Powered Smart Recruitment System

A Flask web app where companies post jobs and candidates apply, then an AI
interviewer ("LUNA") conducts an automated voice interview and scores the
candidate.

## Required Space secrets

Set these under **Settings → Variables and secrets**:

| Secret | Purpose |
| --- | --- |
| `GROQ_API_KEY` | LLM that generates interview questions and chatbot replies |
| `QDRANT_API_URL` | URL of the Qdrant cluster holding the interview questions |
| `QDRANT_API_KEY` | API key for that Qdrant cluster |
| `DATABASE_URL` | *(optional)* Postgres URL; if unset, the app uses SQLite in /tmp |

## Rebuilding the interview-question vector database

The `interview_questions` Qdrant collection is built from
`data/shuffled_questions.json` (4233 Q&A pairs, all-MiniLM-L6-v2, 384-dim,
cosine):

```bash
export QDRANT_API_URL="https://<cluster>.qdrant.io"
export QDRANT_API_KEY="<key>"
python scripts/rebuild_qdrant.py
```