How sajixinc/intellimatch-ai-ats works
Project Type and Purpose
• IntelliMatch-AI-ATS is a full-stack, AI-driven, Streamlit web application (“front-end” + light “back-end” logic in Python) whose primary goal is to help job applicants optimize their resumes by matching them to job descriptions via Google Gemini Pro LLM and to manage email/OTP subscriptions for personalized notifications.
File Structure
• Top-Level Files:
• app.py → main Streamlit entrypoint for resume-matching and all UI flows
• main.py → Streamlit entrypoint for email/OTP subscription features
• final_app.py, email_login.py, email_subscription.py, otp_subscription.py → logical modules encapsulating specific functionality
• assets/ → images used by the UI
• styles/ → CSS overrides (main.css)
• requirements.txt, Procfile, .env (user-created) → environment and deployment config
• No in-project database code; relies on external OTP microservice via HTTP.
README for Insights
• Confirms two big feature groups:
A) Resume-to-Job-Description Matching (PDF processing, LLM calls)
B) Email Subscription & OTP Verification (calls out to external endpoints)
• Lists key libraries: streamlit, PyPDF2, Pillow, python-dotenv, requests.
Main Components
a. User Interface (Streamlit Layer)
– Sub-areas: Resume Upload & Job Description Input; Email/OTP Forms
b. Business Logic Layer (Python Modules)
– Resume Parsing & Matching Service (PyPDF2 + Google Gemini Pro API)
– Keyword Extraction & Suggestions Engine
– Email Subscription Module (email_subscription.py)
– OTP Handling Module (otp_subscription.py)
c. External Services
– Google Gemini Pro API (Generative AI LLM)
– OTP Microservice
• Endpoint1: Otpgenerate (/SubscribeAndFeedback/Otpgenerate/)
• Endpoint2: OtpVerification (/SubscribeAndFeedback/OtpVerfication/)
– Email Delivery Service (inside OTP Microservice)
d. Static Assets
– assets/ (images), styles/main.css
Relationships and Data Flows
• User → Streamlit UI →
• For Resume Matching: calls Resume Parsing → calls Google Gemini Pro → returns analysis → UI displays “Match %,” “Missing Keywords,” “Interview Q&A.”
• For Subscription: posts email to Email Subscription Module → HTTP POST to external Otpgenerate → OTP → user inputs OTP → OTP Handling Module → HTTP POST to OtpVerification → success/failure → UI updates.
[User Browser] → [Streamlit UI] → [Resume Service] → [Google Gemini API]
↘
→ [Email Subscription Module] → [OTP Microservice]
Architectural Patterns
• Layered architecture: UI layer, domain Logic layer, integration layer
• Separation of concerns: each Python module has single responsibility
• External microservice integration rather than in-app database for OTP
Technologies & Frameworks
• Python 3.7+, Streamlit
• PyPDF2 (PDF parsing), Pillow (image handling)
• python-dotenv (.env config), pybase64 (image encoding)
• requests (REST calls to external services)