42jobs — junior job search, done right

42jobs is a job search platform tailored for junior software engineers. It fetches job offers from LinkedIn, filters them with AI for relevance and junior-friendliness, extracts keywords, and generates ATS-optimized CVs — so you spend less time searching and more time landing interviews.

Live at 42jobs.xyz

Table of Contents

• Features
• Screenshots
• Admin Panel
• Tech Stack
• Prerequisites
• Quick Start
• Architecture
• Environment Variables
• Roadmap
• Contributing
• License
• Contact

Features

• Job fetching — pulls offers from LinkedIn via RapidAPI by category
• AI filtering — keeps only relevant offers suitable for junior profiles
• Keyword extraction — identifies technologies, skills, and soft skills per offer
• CV generation — generates ATS-optimized CVs via LLM, customized per job offer
• GitHub import — analyzes your repositories and creates project entries automatically
• Profile management — education, experience, certifications, languages, skills
• Job tracking — status pipeline: saved → CV sent → interview → hired / rejected
• Pluggable AI — swap between Gemini and OpenAI, or add your own provider
• Pluggable job sources — LinkedIn today, more sources planned

Screenshots

Dashboard	Job Search	CV Generation

Tech Stack

Layer	Technology
Backend	.NET 10 (ASP.NET Core Web API), EF Core, JWT
Database	PostgreSQL 16
Frontend	React 19 + React Router 7 + TypeScript (Vite)
AI	OpenAI / Google Gemini (pluggable providers)
Auth	JWT in HttpOnly cookies + BCrypt
Infra	Docker + Docker Compose (dev & prod profiles)
Package mgr	pnpm

Prerequisites

Before you start, make sure you have:

• Docker & Docker Compose
• .NET 10 SDK (optional, only for local dev without Docker)
• Node.js 20+ + pnpm (optional, only for local dev without Docker)
• A RapidAPI account with LinkedIn Jobs API subscription
• An API key for Google Gemini or OpenAI

Quick Start

# 1. Clone
git clone https://github.com/samuelhm/42jobs.git
cd 42jobs

# 2. Configure environment
cp .env.example .env
# Edit .env with your database credentials and a random JWT secret

# 3. Start
make dev-up

The app will be available at:

• Frontend: http://localhost:3000
• API: http://localhost:8080

First-time setup

1. Open the frontend, register an account
2. Go to Admin panel (you'll need to promote your user to Admin in the DB first — see CONTRIBUTING.md)
3. Configure your AI provider (Gemini or OpenAI) with your API key
4. Configure a job provider (LinkedIn RapidAPI) with your API key
5. Create a search category (e.g. "React Developer") and hit Fetch Jobs

Useful commands

make dev-up          # start development services
make dev-down        # stop development services
make dev-restart     # rebuild + restart
make dev-logs        # follow all logs
make prod-up         # start production services
make clean           # stop everything + delete volumes
make release         # create new version tag + trigger deploy

Architecture

42jobs/
├── backend/src/
│   ├── Controllers/     # 13 REST controllers (partial classes, one endpoint per file)
│   ├── Models/          # 22 C# entity models + DTOs
│   ├── Data/            # EF Core DbContext (Fluent API config)
│   ├── Services/
│   │   ├── Ai/          # AI abstraction layer
│   │   │   ├── AiService.cs               # reads prompts from DB, resolves providers
│   │   │   └── Providers/{Gemini,OpenAI}   # low-level API clients
│   │   ├── Jobs/        # Job fetching (background queue via Channel<T>)
│   │   ├── EncryptionService.cs            # API key encryption at rest
│   │   └── JwtService.cs
│   └── Utils/
├── frontend/src/        # React 19 + React Router 7 (Vite + TypeScript)
│   ├── router.tsx        # createBrowserRouter with loaders/actions
│   ├── types/            # Shared TypeScript interfaces
│   ├── utils/            # api, format, match (barrel)
│   ├── hooks/            # useDebounce, usePolling (barrel)
│   ├── context/          # AuthContext, ToastContext (barrel)
│   ├── styles/           # 13 CSS modules by responsibility
│   ├── components/       # Reusable components (7 domain folders + barrel)
│   └── pages/            # Route pages with loaders (4 domain folders + barrel)
├── database/migrations/  # 22 SQL migration + seed files
└── docs/                # Provider guides, encryption docs

AI provider abstraction

Controllers never call AI providers directly. They inject IAiService. The actual provider (Gemini, OpenAI, or future ones) is selected at runtime from the database. Prompts and response schemas are stored in the DB, not hardcoded.

→ Add a new AI provider

Job provider abstraction

Job sources are pluggable. JobFetchService calls all enabled IJobProvider implementations, one per portal. Provider config (keys, URLs) lives in the database and is encrypted at rest.

→ Add a new job provider → How API key encryption works

Admin Panel

The admin panel provides full control over the platform's AI and job provider configuration — no need to touch environment variables or restart containers.

Section	What you can do
AI Services	Enable/disable AI providers (Google Gemini, OpenAI, DeepSeek). Set and encrypt API keys. Mark as free tier (shows a banner to users).
AI Models	Manage available models per provider. Enable/disable specific models (e.g., gemini-2.5-flash, gpt-4o).
AI Prompts	Edit system and user prompts for each AI functionality (filter jobs, extract keywords, dedup, generate CV, parse LinkedIn, analyze GitHub). Set the default model per functionality.
Job Providers	Configure job portals (LinkedIn RapidAPI). Enable/disable, set API keys, base URLs, and provider-specific config as JSON.
CV Templates	Create and manage HTML + CSS templates for CV generation. Templates are selected when regenerating a CV.
Utils	One-click utilities: clean keywords (remove duplicates/normalize), dedup keywords (group synonyms via AI), clean up orphan keywords.
Logs	Audit trail of every AI call and API provider request. Includes timestamps, actor, action, payload, and correlation IDs for tracing.

Environment Variables

Variable	Description	Required
POSTGRES_USER	Database user	Yes
POSTGRES_PASSWORD	Database password	Yes
POSTGRES_DB	Database name	Yes
DATABASE_URL	Full connection string (postgres://user:pass@host:5432/db)	Yes
JWT_SECRET_KEY	Random string for JWT signing (≥ 256 bits)	Yes

AI and job provider API keys are configured via the Admin panel (not env vars) and are encrypted at rest in the database.

Roadmap

• Job fetching from LinkedIn (RapidAPI)
• AI filtering + keyword extraction (Gemini / OpenAI)
• CV generation per job offer
• GitHub project import
• Job tracking pipeline
• API key encryption at rest
• Email notifications for new matching jobs
• More job providers (InfoJobs, Indeed)
• UI tests with Playwright
• Light mode
• Public demo instance

Got an idea? Open an issue or pick one from the roadmap and send a PR!

Contributing

Contributions are welcome! Please read CONTRIBUTING.md for guidelines on how to get started, project conventions, and how to add new providers.

License

Dual-licensed under AGPL v3 for open source / non-commercial use. For commercial use (if you wish to keep your modifications private), contact the author for a commercial license.

Contact

Samuel Hurtado — @hurtadom.dev — samuel@hurtadom.dev

Project Link: https://github.com/samuelhm/42jobs