⚡ NEURAL ATS: AI-Powered Resume Logic Engine

NEURAL ATS is a high-performance, full-stack recruitment tool designed to bridge the gap between static resumes and dynamic job requirements, the application moves beyond simple keyword matching to perform deep semantic "Neural Scans," providing job-seekers with professional-grade ATS compatibility reports.

---

🚀 Core Functionalities

• Dynamic JD Comparison: Analyzes PDF resumes against user-provided Job Descriptions (JD) in real-time.
• Neural Feedback: Generates automated match percentages, identifies critical skill gaps, and provides 3 specific professional improvement directives.
• Cyberpunk Dashboard: A custom-designed, dark-themed UI featuring glassmorphism, neon-glow aesthetics, and responsive layout.
• Real-time Rendering: Utilizes Marked.js to convert raw LLM Markdown responses into clean, human-readable directives.

🛠️ Technical Stack

• Backend: Java 21, Spring Boot 3.2 (MVC Architecture).
• AI Orchestration: Google Gemini 2.5 Flash (Generative AI).
• Binary Processing: Apache PDFBox for robust PDF text extraction.
• Data Handling: Jackson for advanced nested JSON parsing.
• Frontend: Thymeleaf, JavaScript (Fetch API), Bootstrap 5.

🧩 System Architecture

The project follows a modular Controller-Service-Utility pattern to ensure clean separation of concerns and production-ready scalability:

1. Controller Layer: Manages RESTful endpoints and multi-part form data (file + text).
2. Service Layer: Handles business logic and coordinates communication with the Gemini AI model.
3. Utility Layer: Dedicated tools for binary PDF parsing and text sanitization.

⚠️ Challenges & Solutions

During development, we navigated several technical hurdles:

• JSON Parsing Logic: Solved the "undefined" feedback issue by implementing a robust ObjectMapper logic to navigate the nested Gemini API response structure.
• Dynamic Analysis: Shifted from hardcoded keyword matching to a dynamic prompt-based system, allowing for 1:1 matching against any unique Job Description.
• UI/UX: Integrated Marked.js to ensure AI-generated Markdown suggestions were rendered with professional formatting rather than raw text blocks.

🔧 Installation & Setup

1. Clone the repository:

   git clone [https://github.com/shadab0908/NeuralATS.git](https://github.com/shadab0908/NeuralATS.git)

Configure API Key: Add your Gemini API Key to src/main/resources/application.properties:

Properties gemini.api.key=YOUR_API_KEY_HERE Build and Run:

Bash mvn clean install mvn spring-boot:run

Developed by Shadab Javed Mulla as a project to demonstrate Java capabilities and AI integration.