🚀 AutoRef

An AI-Powered, Multi-Tenant Job Outreach Automation Platform

---

📖 Overview

AutoRef is an enterprise-grade, multi-tenant SaaS platform engineered to streamline and automate B2B/Job application outreach. It leverages Google Gemini LLMs to dynamically parse job descriptions and synthesize hyper-personalized referral emails based on the user's resume.

The platform integrates directly with the Gmail API (OAuth 2.0) for email dispatch and employs an intelligent event-driven background scheduler (APScheduler) for automated, multi-stage follow-up sequences and active reply detection.

This project demonstrates strong proficiency in Systems Architecture, Secure Authentication, Third-Party Integrations, and Asynchronous Processing.

---

🏗️ System Architecture

AutoRef is built on a modern decoupled architecture, ensuring scalability, maintainability, and a clear separation of concerns between the presentation and data/processing layers.

graph TD
    Client("💻 Next.js Client<br/>(React / Tailwind)")
    
    subgraph FastAPI Backend
        Auth("🔐 Auth Router<br/>(JWT / Google ID)")
        API("⚡ REST API<br/>(Protected Endpoints)")
        Scheduler("⏱️ APScheduler<br/>(Background Jobs)")
        ORM("🗄️ SQLAlchemy<br/>(Data Access)")
    end

    subgraph External Services
        Gemini("🧠 Google Gemini<br/>(LLM Engine)")
        Gmail("📧 Gmail API<br/>(OAuth 2.0)")
        GoogleAuth("👤 Google Identity<br/>(Sign-In)")
    end
    
    DB[("SQLite<br/>Database")]

    Client <-->|Google ID Token| GoogleAuth
    Client <-->|Login / Register| Auth
    Client <-->|Bearer JWT| API
    
    Auth <--> ORM
    API <--> ORM
    Scheduler <--> ORM
    ORM <--> DB
    
    API <-->|Prompt / Parse| Gemini
    API <-->|Send Drafts| Gmail
    
    Scheduler -.->|Auto-Send Follow-ups| Gmail
    Scheduler -.->|Poll Inbox Replies| Gmail

Loading

---

✨ Core Features & Technical Highlights

🛡️ Enterprise-Grade Auth & Multi-Tenancy

• Robust Security: Implements JWT-based authentication with bcrypt password hashing and Google Sign-In (OAuth ID Tokens).
• 3-Tier Dependency Injection: FastAPI routes are protected via a strictly typed dependency chain (Authenticated → Approved → Admin).
• Admin Gatekeeping: First-user bootstrap strategy auto-approves the admin; subsequent registrations enter a pending state until manually approved via the Admin Panel, strictly preventing platform abuse.
• Row-Level Isolation: 100% data isolation across tenants using strict user_id foreign key scoping on all ORM queries.

🧠 Context-Aware AI Generation (Gemini)

• Semantic Parsing: Dynamically extracts Company, Role, and Skills from raw Job Description URLs/text.
• Role-Specific Prompt Engineering: Utilizes structured role_configs (Backend/SDE, Fintech, Data Engineering) to instruct the LLM on which specific achievements to highlight from the user's profile, generating high-converting B2B copy.

⏱️ Asynchronous Workflow Engine

• Intelligent Follow-ups: APScheduler orchestrates a stateful, multi-stage follow-up pipeline.
• Rate-Limit Respecting: Jobs are naturally throttled (1 per minute) to respect Gmail API limits and prevent spam flagging.
• Idempotent Processing: Terminal states (sent, failed, cancelled) ensure that transient network failures do not result in duplicate emails to recruiters.

🔄 Automated Inbox Monitoring

• Event-Driven Polling: Periodically monitors the connected Gmail inbox.
• Reply Detection: Autonomously detects replies from external domains, updates the application thread state, and immediately halts pending follow-up jobs.

---

🛠️ Technology Stack

Domain	Technologies
Frontend	Next.js 16, React, Tailwind CSS, Context API
Backend	FastAPI, Python 3, Pydantic, Passlib, python-jose
Database	SQLite, SQLAlchemy (ORM), Alembic
Background Jobs	APScheduler
Integrations	Google Gemini API, Gmail API, Google Sheets API

---

🚀 Local Development Setup

To run AutoRef locally, spin up both the FastAPI backend and the Next.js frontend.

Prerequisites

• Python 3.9+
• Node.js 18+
• Google Cloud Console Account (Gmail API enabled, OAuth 2.0 Credentials configured)
• Google Gemini API Key

1. Backend Initialization (FastAPI)

cd backend
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

# Configure Environment
cp .env.example .env
# --> Edit .env with your specific API Keys and JWT secrets

# Start the API Server (Auto-migrates DB on first run)
uvicorn main:app --reload --port 8000

2. Frontend Initialization (Next.js)

Open a new terminal window:

cd frontend
npm install

# Configure Environment
cp .env.example .env.local
# --> Edit .env.local with your backend URL and Google Client ID

# Start the Development Server
npm run dev

Navigate to http://localhost:3000 to access the application.

---

📸 Application Previews

Dashboard Analytics	Outreach Generator	Admin Access Control
		(Admin panel for user approval routing)

---

📄 Author Resumes

Target Engineering Domain	Resume Link
SDE / Backend Engineering	View PDF (Google Drive)

---

Architected for scale, built for speed.