Job Hunt Duo

AI-powered job search platform — University Project (Metode de Dezvoltare Software, Univ. of Bucharest)

---

Overview

Job Hunt Duo uses two AI agents to automate and personalise the job search:

Agent	Model	Role
Agent 1 – Job Finder Bot	qwen2.5-1.5b-instruct via LMStudio	Searches Adzuna, RemoteOK, Arbeitnow & HN Who's Hiring, scores each listing against your CV (Epic 4)
Agent 2 – CV Tailor Bot	llama-3.2-3b-instruct via LMStudio	Rewrites your CV for a specific job and generates cover letters (Epic 5)

---

Technology Stack

Layer	Technology
Backend	Django 4.2 + Django REST Framework
Auth	JWT (djangorestframework-simplejwt)
Database	SQLite (dev) / PostgreSQL 15 (Docker/prod)
Frontend	React 18 + React Router v7
Styling	Tailwind CSS + shadcn/ui
AI Agents	LMStudio (OpenAI-compatible local API)
Containerisation	Docker + Docker Compose
CI	GitHub Actions (see .github/workflows/ci.yml)

---

Quick Start (Local Development)

Prerequisites

• Python 3.11+
• Node.js 20+
• LMStudio running on localhost:1234 with both models loaded

Backend

cd backend
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver 0.0.0.0:8001

Frontend

cd frontend
yarn install
yarn start          # runs on localhost:3000

Docker (full stack with PostgreSQL)

cp docker/.env.example docker/.env
# Fill in POSTGRES_PASSWORD and SECRET_KEY
docker compose -f docker/docker-compose.yml up --build

---

Environment Variables

Variable	Default	Description
SECRET_KEY	(insecure dev key)	Django secret key — change in production
DATABASE_URL	SQLite	PostgreSQL URL for Docker/prod
DEBUG	True	Set to False in production
LMSTUDIO_BASE_URL	http://localhost:1234/v1	LMStudio local API base URL
JOB_FINDER_MODEL	qwen2.5-1.5b-instruct	Model used by Agent 1
CV_TAILOR_MODEL	llama-3.2-3b-instruct	Model used by Agent 2

---

API Reference

All routes are prefixed with /api.

Authentication (Epic 1)

Method	Endpoint	Description
POST	/api/auth/register/	US-01 Register account
POST	/api/auth/login/	US-02 Login (returns JWT)
POST	/api/auth/refresh/	Refresh access token
GET	/api/auth/me/	Get current user

CV Management (Epic 2)

Method	Endpoint	Description
POST	/api/cv/upload/	US-03 Upload PDF, auto-parse
GET	/api/cv/profile/	Get parsed CV data
PATCH	/api/cv/profile/update/	US-04 Edit CV fields

Preferences (Epic 3)

Method	Endpoint	Description
GET/PUT	/api/preferences/	US-05 Job preferences
GET	/api/preferences/sources/	US-06 List job sources
PATCH	/api/preferences/sources/<source>/toggle/	Toggle source on/off

---

Project Structure

job-hunt-duo/
├── backend/
│   ├── server.py              # ASGI entry point (Django via uvicorn)
│   ├── manage.py
│   ├── jobhunt/               # Django project config
│   │   ├── settings.py
│   │   └── urls.py
│   ├── apps/
│   │   ├── users/             # Epic 1 – Auth
│   │   ├── cv/                # Epic 2 – CV Upload & Parsing
│   │   └── preferences/       # Epic 3 – Preferences & Sources
│   ├── agents/
│   │   ├── lmstudio_client.py # Shared LMStudio HTTP client
│   │   ├── job_finder_stub.py # Agent 1 stub (Epic 4)
│   │   └── cv_tailor_stub.py  # Agent 2 stub (Epic 5)
│   └── tests/
│       └── test_api.py        # pytest test suite (15 tests, all passing)
├── frontend/
│   └── src/
│       ├── pages/             # Auth, Dashboard, CV, Preferences
│       ├── components/        # Sidebar, TopBar, layout
│       ├── context/           # AuthContext (JWT management)
│       └── services/          # Axios API client
├── docker/
│   ├── docker-compose.yml
│   ├── Dockerfile.backend
│   ├── Dockerfile.frontend
│   └── .env.example
├── README.md
└── AI_REPORT.md

---

Running Tests

Backend (pytest)

cd backend && pytest tests/ -v

Frontend

cd frontend && yarn test

---

Backlog & User Stories

Full backlog: job_hunt_duo_backlog_v2.html

Implemented (Sprint 1):

• US-01 User registration
• US-02 User login (JWT)
• US-03 CV PDF upload + auto-parsing
• US-04 Edit extracted CV data
• US-05 Job preferences (title, location, work type, seniority)
• US-06 Job source configuration (Adzuna, RemoteOK, Arbeitnow, HN)

Upcoming:

• US-07–09 Agent 1: Job Finder Bot (Epic 4)
• US-10–12 Agent 2: CV Tailor Bot (Epic 5)
• US-13–15 Email digest, tracking, mobile UX (Epic 6)

---

Architecture Diagrams

TODO (required for submission): Create in draw.io and link below.

1. Component Architecture Diagram (Docker services)
2. UML Use Case Diagram (Job Seeker actor)
3. AI Agent Flow Diagram (Agent 1 + Agent 2)

---

AI Usage Report

See AI_REPORT.md

---

License

University project — University of Bucharest, 2025.