JobTrackerPro 🚀

"I stopped using Excel to track my job applications. I architected an Enterprise-Grade platform instead."

JobTrackerPro is a robust, full-stack solution designed to visualize, automate, and manage the interview pipeline. Unlike simple CRUD tutorials, this project demonstrates Advanced System Design, Cloud-Native Architecture, and Production-Grade Security.

---

Who Is This For?

• Job seekers who want zero-effort tracking via email automation
• Engineers interested in real-world AI ingestion pipelines
• Developers learning enterprise Spring Boot + Angular architecture
• Open-source contributors looking for a non-trivial system to extend

---

📌 Project Status

• JobTrackerPro is actively developed and open for contributions.
• The dev branch is the primary development branch.
• The main branch is reserved for stable, production-ready releases.

---

🔗 Quick Links

• 🚀 Live Demo: https://jobtrackerpro.in
• 📂 Source Code: https://github.com/thughari/JobTrackerPro

---

🏗️ System Architecture & App Flow

JobTrackerPro is built as a highly automated, decoupled system focusing on security, latency reduction, and seamless user experiences.

🔄 End-to-End Application Flow

flowchart TD
    %% Styling
    classDef service fill:#4A90E2,stroke:#357ABD,stroke-width:1.5px,color:#fff;
    classDef utility fill:#50E3C2,stroke:#34A790,stroke-width:1.5px,color:#333;
    classDef storage fill:#F5A623,stroke:#C68015,stroke-width:1.5px,color:#fff;
    classDef trigger fill:#D0021B,stroke:#9E0010,stroke-width:1.5px,color:#fff;

    A[Gmail Webhook / Push Notice]:::trigger --> B(GmailWebhookService):::service
    B --> C{SmartExtractionService}:::service
    
    %% Template Ingest Pathway
    C -->|LinkedIn / Indeed Template| D(TemplateParser):::utility
    D -->|Extract Role & Company| E{Body Template Matches?}
    E -->|Yes| F[Extract from Body & Clean Location]:::utility
    E -->|No| G[Extract from Subject & Find Location]:::utility
    D -->|Extract & Clean URL| H[UrlParser: Decode next= Parameter]:::utility
    D -->|Determine Status| I[Strip Safety Warnings & Scan Keywords]:::utility
    
    %% AI Ingest Pathway
    C -->|Unmatched Emails| J{app.gemini.enabled?}
    J -->|true| K(GeminiExtractionService):::service
    J -->|false| L(MockGeminiService):::service
    
    %% Job Matching & Storage
    F --> M(JobService: createOrUpdateJob):::service
    G --> M
    K --> M
    L --> M
    
    M --> N{findBestMatch}:::service
    N -->|Stricter Token Matching for Short Names| O[Update Existing Job & Append Notes]:::storage
    N -->|No Active Match| P[Create New Job Entry]:::storage
    
    O --> Q[(Database: PostgreSQL)]:::storage
    P --> Q
    Q --> R(Caffeine Cache Eviction / Angular Signal Update):::service

The system operates across four key pipeline phases:

1. 📬 Automated Ingestion & Webhook Layer

• Gmail Sync Integration: Users link their Gmail account via OAuth2. A Google Pub/Sub topic listens to mailbox events and sends real-time push notifications to GmailWebhookService.
• Smart routing (SmartExtractionService): Ingested emails are routed through a @Primary interceptor. It checks if the email is a standard template from LinkedIn or Indeed.
  • Matched: Processed locally instantly (0ms AI cost, sub-millisecond execution).
  • Unmatched: Delegated to Google Gemini 2.0 Flash (or local MockGeminiService if offline).

2. ⚙️ Manual Extraction Engine (TemplateParser)

• Forward Header Reconstruction: Recognizes forwarded email structures so you can forward confirmation emails directly to your sync mailbox.
• Indeed Body Parser: Extracts the role, company name, and location directly from standard Indeed body layouts to avoid misidentifying hyphenated roles (e.g. Java Backend Developer - MS parses Java Backend Developer - MS as the role and Capco as the company).
• Url query decoding: Recognizes apply.indeed.com confirmation links, extracts their next= query parameter, URL-decodes it, and stores the direct public listing URL (e.g., https://in.indeed.com/viewjob?jk=...), reducing length from 220+ to ~50 characters.
• Location & Status Cleaning: Removes reviews meta-data (e.g. Remote 95 reviews -> Remote) and filters safety warning disclaimers to prevent false matches (e.g., preventing "without an interview" from triggering an Interview Scheduled status).

3. 🧠 Smart Deduplication & Matching (JobService)

• When a parsed job hits JobService.createOrUpdateJob(), it compares the company and role with your active board entries:
  • Strict Token matching: If either company name is $\le 3$ characters (e.g., MS), it requires an exact word-token match rather than a simple substring contains check (preventing MS from matching ORION SYSTEMS).
  • Status Updates: If a match is found (e.g. an interview invite email for a job you already applied to), it updates the existing entry in-place and appends the email details to the transaction notes. Otherwise, it inserts a new job entry.

4. 🎨 Reactive UI & In-Memory Caching

• State Management: The Angular frontend uses Angular Signals for reactive state updates, giving the user instant UI response.
• Caffeine In-Memory Caching: User profiles, dashboard analytics, and job lists are cached on the Spring Boot backend, automatically evicted upon job modifications to keep the UI up-to-date while protecting the DB from heavy read traffic.
• Don't-Drown Slice Scaling: The D3.js Donut Chart scales tiny slices (e.g., 1 application status out of 400+) up to a minimum 2.5% display size so they are visible and hoverable, while keeping tooltip values mathematically accurate.

---

🛠️ The Tech Stack

Domain	Technology	Key Usage
Backend	Java 21	Modern JVM features (Records, Pattern Matching)
Framework	Spring Boot 3.4	REST API, Security, Data JPA
Database	PostgreSQL	Supabase managed instance (Transaction Mode)
AI Model	Gemini 2.0 Flash	Intelligent email parsing
Storage	Cloudflare R2	S3-compatible object storage
Frontend	Angular 17	Signals, Standalone Components, Optimistic UI
Styling	TailwindCSS	Utility-first styling, Dark Mode
DevOps	Docker & Cloud Run	Containerized serverless deployment

---

📂 Project Structure

The repository is structured as a Monorepo:

JobTrackerPro/
├── backend/                # Spring Boot API
│   ├── src/main/java/      # Controllers, Services, DTOs
│   ├── src/main/resources/ # Configurations for local, dev, prod # use local for dev
│   ├── Dockerfile          # Backend Container Config
│   ├── scripts             # Utility Scripts (e.g., simulate-email.sh)
│   └── service.yaml        # Google Cloud Run Config
├── frontend/               # Angular UI
│   ├── src/app/            # Components, Services, Guards
│   └── tailwind.config     # CSS Configuration
├── .github/workflows/      # CI/CD (GCP Cloud Run & GitHub Pages)
├── docker-compose.yml      # Local Dev Infrastructure 
└── README.md               # Documentation

---

🚀 One-Command Onboarding

You can run the entire ecosystem locally with zero configuration. The app automatically uses Mock AI and Local File Storage so you don't need any paid API keys to start contributing.

1. Spin up Infrastructure

Requires Docker Desktop. This starts PostgreSQL and MailHog (Email Trap).

docker-compose up -d

2. Start the Backend

cd backend
./mvnw spring-boot:run

3. Start the Frontend

cd frontend
npm install
npm start

4. Test the Features

• Dashboard: Access at http://localhost:4200.
• Emails: View outgoing emails at http://localhost:8025 (MailHog).
• AI Ingestion: Use the scripts/simulate-email.sh to see the Mock AI create jobs automatically.

---

📸 Screenshots

Interactive Dashboard	Profile & Automation
Real-time D3.js analytics and charts	Email forwarding setup and secure settings

---

🤝 Contributing

Contributions are welcome!

1. Fork the repository.
2. Create a Feature Branch (git checkout -b feature/AmazingFeature).
3. Commit your changes.
4. Push to the branch.
5. Open a Pull Request to branch dev.

---

📄 License

This project is licensed under the MIT License.

---