CareerOS — Your AI Career Operating System

CareerOS is an AI-powered career operating system that automates job search, interview preparation, resume optimization, and networking. It orchestrates 8 specialized agents coordinated by a central planner to provide personalized career coaching at scale.

Quick Start

# Frontend
npm install
npm run dev               # → http://localhost:8080

# Backend
cd backend
pip install -r requirements.txt
alembic upgrade head
uvicorn app.main:app --reload  # → http://localhost:8000

Or use Docker: docker compose up (Postgres, Redis, Qdrant, backend).

Tech Stack

Layer	Technology
Frontend	React 18 + Vite + TypeScript + Tailwind CSS
UI	shadcn/ui (Radix primitives), Framer Motion, Recharts
State	TanStack React Query, React Router DOM
Backend	Python FastAPI + SQLAlchemy 2.0 (async) + Alembic
Database	PostgreSQL 16
Vector DB	Qdrant (semantic memory / RAG)
Cache / Queue	Redis + RQ worker
AI Agents	LangGraph + LangChain
LLM Providers	OpenAI, Anthropic, Google Gemini, Groq, DeepSeek, Mistral, Ollama (fallback chain)
Search	Tavily, Google CSE, Brave, SerpAPI, Exa, SearXNG (adapter chain)
Auth	Firebase (email/password + Google SSO) with JWT
Container	Docker Compose (4 services)
Deployment	Vercel (frontend), Render (backend)

Architecture

Browser → Vite/React → REST/WS → FastAPI → Agents (LangGraph)
                                          → PostgreSQL (canonical data)
                                          → Qdrant (vector memory)
                                          → Redis (cache, queues)
                                          → LLM providers (fallback chain)

8 agents orchestrated via LangGraph:

1. Planner — career strategy and goal breakdown
2. Job Agent — full-time job discovery and matching
3. Internship Agent — internship, hackathon, fellowship scanning
4. Research Agent — company intelligence and market analysis
5. Resume Agent — ATS analysis and tailored rewrites
6. Interview Agent — mock interviews with scoring and feedback
7. Networking Agent — contact discovery and outreach generation
8. Opportunity Monitor — 24/7 background scanning and alerts

A memory layer (Qdrant + PostgreSQL) stores user preferences, application history, interview feedback, and career goals so recommendations improve over time.

Project Structure

├── src/                 # React frontend (Vite)
│   ├── pages/           # 14+ routed pages
│   ├── components/      # shadcn/ui + custom components
│   └── lib/             # API client, auth, utilities
├── backend/
│   ├── app/
│   │   ├── main.py      # FastAPI entry point
│   │   ├── api/v1/      # REST routes (15+ modules)
│   │   ├── agents/      # LangGraph agent implementations
│   │   ├── models/      # SQLAlchemy ORM models
│   │   ├── services/    # Business logic
│   │   ├── memory/      # Qdrant vector store client
│   │   ├── search/      # Multi-provider search adapters
│   │   └── tasks/       # Background workers
│   ├── tests/           # 300+ test functions
│   └── alembic/         # Database migrations
├── docker-compose.yml   # Local dev services
├── Dockerfile.frontend  # Frontend container
└── backend/Dockerfile   # Backend container

Key Features

• ATS Resume Analysis — keyword gap detection, tailored rewriting
• Mock Interviews — AI-powered with real-time scoring (STAR, technical, system design)
• Opportunity Discovery — scans 50+ sources with 78% match accuracy
• Networking Automation — contact discovery + personalized outreach
• Pipeline Tracking — Kanban-style application management
• Career Analytics — conversion tracking, skill demand insights
• Persistent Memory — learns from every interaction across sessions

Deployment

• Frontend: Deploy dist/ to Vercel (SPA rewrite via vercel.json)
• Backend: Deploy to Render using render.yaml (Python)
• Container: docker compose -f docker-compose.prod.yml up

License

MIT