Crewmap

Agentic Job Roadmap Generator

Crewmap converts real job listings into structured, role-specific learning roadmaps using an agentic AI system.

Live Demo

Crewmap is an end-to-end system that takes a job listing (such as a LinkedIn job URL) and returns a clear, actionable learning roadmap tailored to that role.

It is designed as a system with emphasis on correctness, performance, concurrency safety, and cost-efficient AI usage.

• Job-specific roadmap generation
• Agent-based AI reasoning with clear separation of concerns
• Redis caching and distributed locking
• Stateless, horizontally scalable backend
• Modern React frontend with local history
• Export roadmaps as Markdown or PDF

Frontend

• React + TypeScript
• Tailwind CSS
• LocalStorage for history
• Collapsible and interactive roadmap UI

Backend

• FastAPI (Python)
• Redis for caching and locks
• Stateless API design

AI Layer

• CrewAI orchestration

• Two agents:

  • Job Analysis Agent (uses Serper for search context)
  • Roadmap Generation Agent

• Gemini as the LLM provider

The AI system is split into two agents:

1. Job Agent

   • Analyzes the job description
   • Uses Serper to fetch company and role context
   • Condenses relevant information to reduce token usage and latency

2. Roadmap Agent

   • Receives enriched job context
   • Focuses on generating a structured learning roadmap
   • Produces phased or time-based learning guidance

This design improves speed, reduces cost, and keeps agent responsibilities clearly separated.

1. User submits a job URL via the frontend

2. Backend extracts job ID and checks Redis cache

3. If cached, the roadmap is returned immediately

4. If not cached:

   • A Redis lock is acquired
   • The AI workflow is executed
   • The result is cached

5. The roadmap is returned to the frontend for rendering and export

Frontend

• React 18+
• TypeScript
• Tailwind CSS

Backend

• Python 3.11+
• FastAPI
• Pydantic
• Redis

AI

• CrewAI
• Gemini
• Serper

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt
uvicorn roadmap.api:app --reload

• API: http://localhost:8000
• Docs: http://localhost:8000/docs

npm install
npm run dev

• App: http://localhost:5173

crewmap/
├── frontend/        # React frontend
├── roadmap/         # Backend service
│   ├── api.py
│   ├── crew.py
│   ├── services/
│   ├── cache/
│   └── utils/
└── README.md

• Developers preparing for specific roles
• Students planning structured learning paths
• Career switchers entering new domains
• Engineers exploring agentic AI system design

Contributions are welcome.

• Bug fixes and improvements
• Feature additions
• UI/UX enhancements
• Agent logic and prompt improvements
• Documentation updates

PRs welcome!