Wand 🪄

Wand is an intelligent AI-powered career assistant that helps users analyze job postings, optimize their resumes, and generate tailored cover letters. It leverages advanced LLMs (Gemini 2.5 Pro & Flash) to provide deep insights and actionable feedback.

🚀 Features

• Job Analysis: deeply analyzes job descriptions to extract requirements, skills, and expected salary.
• Resume Parsing: Extracts structured data from PDF resumes using AI.
• Gap Analysis: Compares your resume against job requirements to identify missing qualifications and keywords.
• Resume Optimization: Provides specific "Add", "Update", or "Delete" suggestions to tailor your resume.
• Cover Letter Generation: improved cover letters (Professional, Creative, Storytelling modes) tailored to the specific job.
• Application Tracking: Kanban-style board to track job applications.

🏗️ Architecture

The project is organized into three main components:

1. Frontend (/frontend)

The user interface built with Next.js 14, TypeScript, and Tailwind CSS.

• Tech Stack: Next.js, React, Zustand (State Management), Framer Motion (Animations).
• Key Functionality: Job dashboard, interactive resume editor, analysis visualization, real-time updates via WebSockets.

2. API (/api)

The backend REST API built with FastAPI.

• Tech Stack: FastAPI, SQLAlchemy (SQLite), Pydantic.
• Key Functionality: Data persistence, request orchestration, file handling, and communication with the Engine.
• Workers: Uses Celery (with Redis) for background processing of long-running analysis tasks.

3. Engine (/engine)

The intelligence core of the application.

• Tech Stack: Python, Inspector (Structured Output), PyMuPDF.
• AI Models:
  • Gemini 2.5 Pro: Used for complex reasoning (Qualification Validation, Cover Letter Writing).
  • Gemini 2.5 Flash: Used for high-speed tasks (Parsing, Formatting Checks, Keyword Matching).
• Modules:
  • job/parser: Extracts structured job data.
  • analysis/qualification_check: validatates resume claims against requirements.
  • analysis/formatting_check: Reviews resume formatting and chronology.
  • cover_letter/generator: Generates personalized cover letters.

🛠️ Setup & Installation

Prerequisites

• Python 3.10+
• Node.js 18+
• Redis (for background tasks)
• Google Gemini API Key

Backend Setup

1. Navigate to root directory.
2. Create virtual environment:

   python -m venv .venv
   source .venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set environment variables in .env:

   GEMINI_API_KEY=your_key_here
   REDIS_URL=redis://localhost:6379/0

5. Start the server:

   uvicorn api.main:app --reload

Frontend Setup

1. Navigate to frontend:

   cd frontend

2. Install dependencies:

   npm install

3. Start dev server:

   npm run dev

🧠 AI Model Configuration

The system uses a hybrid model approach for optimal performance:

• Gemini 2.5 Pro: used for logic-heavy tasks requiring "reasoning" capabilities.
• Gemini 2.5 Flash: Used for speed-critical extraction and matching tasks.

📄 License

MIT