AI Job Agent

An AI-powered job discovery and application assistant that helps users discover relevant software and AI/ML opportunities, analyze job fit using LLMs, manage resumes, and track applications from a single platform.

🚀 Overview

AI Job Agent is a full-stack AI application designed to simplify the job-search process.

The platform aggregates jobs from multiple hiring platforms, allows users to search and filter opportunities, and provides AI-powered analysis to help candidates understand their suitability for a particular role.

The application also provides user authentication, profile management, resume management, and application tracking.

✨ Features

🔐 Authentication

Authentication is handled using Supabase Auth.

Supported authentication methods:

Email/password authentication
Email confirmation
Google OAuth
Persistent authentication sessions
Secure sign-out

Email confirmation emails are configured through Gmail SMTP using a Google App Password.

📧 Email System

The application uses:

Supabase Auth for authentication
Gmail SMTP for transactional authentication emails
Email confirmation for newly created accounts

The email flow is:

Create Account
      ↓
Supabase Auth
      ↓
Gmail SMTP
      ↓
Confirmation Email
      ↓
User confirms email
      ↓
Account activated

Gmail SMTP configuration uses:

SMTP Host: smtp.gmail.com
SMTP Port: 465
Authentication: Gmail App Password

The Gmail App Password is stored securely in the Supabase SMTP configuration and is never stored in the source code.

💼 Job Discovery

AI Job Agent aggregates job opportunities from multiple hiring platforms.

Supported Sources

Greenhouse
Lever
Ashby
Workable

Users can search jobs using:

Job role
Location
Seniority
Remote-only preference

The architecture allows additional ATS/job-board integrations to be added independently.

🤖 AI-Powered Job Matching

The application uses Groq for fast LLM inference.

The current job-fit analysis uses:

Groq
  ↓
openai/gpt-oss-120b

AI analysis can consider:

Job description
Required skills
Candidate skills
Candidate experience
Technical background
Matching qualifications
Missing qualifications

The analysis is triggered when the user chooses to calculate their fit for a job.

Job Fit Pipeline

Job Description
       ↓
Job Parser
       ↓
Structured Job Data
       ↓
Candidate Profile
       ↓
Resume Information
       ↓
Groq LLM
       ↓
AI Job Fit Analysis
       ↓
Match Explanation

The application does not run an expensive LLM analysis for every job during the initial search. AI analysis is performed when the user actually requests a fit analysis.

🧠 Embeddings & Semantic Matching

The backend also contains an embedding service for semantic job/candidate matching.

The architecture supports embedding-based retrieval where job descriptions and candidate information can be converted into vector representations and compared based on semantic similarity.

This allows the system to move beyond simple keyword matching.

Planned/ongoing improvements include:

Semantic job retrieval
Resume-job similarity
Skill-gap analysis
Candidate ranking based on relevance
Hybrid retrieval + LLM reasoning

📄 Resume Management

The resume system is designed to support:

Resume upload
Resume parsing
Resume information extraction
Structured resume data
Resume management
Resume-job matching

The resume information can later be used by the AI job-matching pipeline.

Resume Pipeline

Resume Upload
      ↓
Resume Parser
      ↓
Extract Resume Information
      ↓
Structured Resume
      ↓
Candidate Profile
      ↓
Job Matching

👤 Profile Management

Users can manage their personal information through the profile section.

Profile information includes:

First name
Last name
Phone number
Date of birth
Gender
Profile photo
Education
Skills
Experience
Job preferences
Resume

The profile is associated with the authenticated Supabase user.

📋 Application Tracking

The application includes an application-tracking system for managing job applications.

Supported/planned application states include:

Saved
  ↓
Applied
  ↓
Interview
  ↓
Offer / Rejected

Users can maintain application history and track the status of opportunities they have applied to.

🎨 User Interface

The frontend is built with Streamlit and customized using HTML/CSS.

The application includes:

Modern navigation bar
Profile dropdown
Floating profile card
Light/dark theme
Responsive layout
Job search interface
Job cards
Profile management
Resume management
Application tracking

The UI is designed around a clean, minimal interface inspired by modern job-search platforms.

🏗️ Architecture

                         AI JOB AGENT
                              │
                ┌─────────────┴─────────────┐
                │                           │
             Frontend                    Backend
                │                           │
           Streamlit                  Python Services
                │                           │
                │              ┌────────────┼────────────┐
                │              │            │            │
                │          Job Sources   AI/LLM      Database
                │              │            │            │
                │              │          Groq       Supabase
                │              │            │
                │              │       Job Scoring
                │              │
                │        ┌─────┴─────┐
                │        │           │
                │    Greenhouse    Lever
                │    Ashby         Workable
                │
                └────────────────────────────────────────

🛠️ Tech Stack

Frontend

Python
Streamlit
HTML
CSS
JavaScript where required

Backend

Python
REST APIs
ATS integrations
Modular service architecture

Authentication

Supabase Auth
Google OAuth
Email/password authentication
Gmail SMTP
Google App Passwords

Database

Supabase
PostgreSQL
Row Level Security (RLS)

AI / ML

Groq
openai/gpt-oss-120b
Embeddings
NLP
Resume parsing
Structured job analysis
Semantic matching

🗄️ Database

The application uses Supabase PostgreSQL for persistent data.

Current tables include:

profiles
resumes
jobs
saved_jobs
applications
job_fit_analyses
recent_searches

Row Level Security is enabled to protect user-specific data.

🔄 Job Aggregation Pipeline

                         User Search
                              │
                              ▼
                     Job Aggregator
                              │
             ┌────────────────┼────────────────┐
             │                │                │
             ▼                ▼                ▼
        Greenhouse          Lever            Ashby
             │                │                │
             └────────────────┼────────────────┘
                              │
                              ▼
                          Workable
                              │
                              ▼
                     Normalize Jobs
                              │
                              ▼
                       Filter Jobs
                              │
                              ▼
                       Search Results
                              │
                              ▼
                         User

🤖 AI Job Fit Pipeline

                    Job Description
                           │
                           ▼
                      Job Parser
                           │
                           ▼
                  Structured Job Data
                           │
                           │
             ┌─────────────┴─────────────┐
             │                           │
             ▼                           ▼
      Candidate Profile               Resume
             │                           │
             └─────────────┬─────────────┘
                           │
                           ▼
                       Groq LLM
                  openai/gpt-oss-120b
                           │
                           ▼
                   Job Fit Analysis
                           │
                           ▼
                 Match Explanation

🔐 Authentication Architecture

Email Authentication

User
 │
 ▼
Create Account
 │
 ▼
Supabase Auth
 │
 ▼
Gmail SMTP
 │
 ▼
Confirmation Email
 │
 ▼
Email Confirmation
 │
 ▼
Authenticated User

Google OAuth

User
 │
 ▼
Continue with Google
 │
 ▼
Google OAuth
 │
 ▼
Supabase Callback
 │
 ▼
Authorization Code
 │
 ▼
Session Exchange
 │
 ▼
Authenticated Application

📁 Project Structure

AI-Job-Agent/
│
├── backend/
│   ├── models/
│   │   ├── job.py
│   │   ├── resume.py
│   │   └── structured_job.py
│   │
│   ├── services/
│   │   ├── auth_service.py
│   │   ├── embedding_service.py
│   │   ├── job_aggregator.py
│   │   ├── job_filter.py
│   │   ├── job_parser.py
│   │   ├── job_scorer.py
│   │   ├── llm_service.py
│   │   ├── resume_parser.py
│   │   └── supabase_service.py
│   │
│   └── sources/
│       ├── ashby.py
│       ├── base.py
│       ├── greenhouse.py
│       ├── lever.py
│       └── workable.py
│
├── frontend/
│   ├── components/
│   │   └── navbar.py
│   │
│   ├── styles/
│   │   └── theme.py
│   │
│   ├── views/
│   │   ├── applications.py
│   │   ├── auth.py
│   │   ├── jobs.py
│   │   ├── profile.py
│   │   └── resume.py
│   │
│   └── streamlit_app.py
│
├── tests/
│
├── .env
├── .gitignore
├── requirements.txt
└── README.md

⚙️ Installation

1. Clone the repository

git clone https://github.com/YOUR_USERNAME/AI-Job-Agent.git
cd AI-Job-Agent

2. Create a virtual environment

Windows:

python -m venv job-agent-env

Activate it:

.\job-agent-env\Scripts\Activate.ps1

3. Install dependencies

pip install -r requirements.txt

4. Configure environment variables

Create a .env file in the project root.

Example:

SUPABASE_URL=your_supabase_project_url
SUPABASE_KEY=your_supabase_publishable_key
GROQ_API_KEY=your_groq_api_key

Additional environment variables may be required as the application grows.

5. Run the application

streamlit run frontend\streamlit_app.py

The application will normally be available at:

http://localhost:8501

🔑 Environment & Security

Never commit credentials to GitHub.

The following must remain private:

.env
Supabase credentials
Groq API key
Gmail App Password
OAuth secrets
Database credentials

Recommended .gitignore entries:

.env
.env.*
!.env.example

__pycache__/
*.py[cod]

.venv/
venv/
job-agent-env/

.streamlit/secrets.toml

🧪 Testing

Tests are located in:

tests/

Run all tests:

pytest

Run a specific test:

pytest tests/test_greenhouse.py

🧩 Backend Services

`auth_service.py`

Handles:

Sign up
Sign in
Sign out
Google OAuth
Session management

`supabase_service.py`

Handles Supabase client initialization and authentication session storage.

`job_aggregator.py`

Coordinates multiple job sources and combines their results.

`job_filter.py`

Filters jobs according to search criteria.

`job_parser.py`

Extracts structured information from job descriptions.

`job_scorer.py`

Handles AI-assisted job-fit analysis.

`llm_service.py`

Provides the LLM interface used by the application.

`embedding_service.py`

Provides embedding functionality for semantic matching.

`resume_parser.py`

Processes uploaded resumes and extracts relevant information.

🗺️ Roadmap

Phase 1 — Foundation

Phase 2 — Job Discovery

Phase 3 — Candidate Profile

Phase 4 — Resume Intelligence

Phase 5 — AI Matching

Phase 6 — Applications

Phase 7 — AI Application Assistant

Tailored resume generation
Cover-letter generation
Application preparation
Browser-based application assistance
Playwright automation
Manual fallback for unsupported application flows

Phase 8 — Deployment

🚧 Project Status

AI Job Agent is currently under active development.

The core foundation includes:

Supabase authentication
Google OAuth
Email confirmation
Gmail SMTP
Supabase PostgreSQL
Row Level Security
Multiple ATS integrations
Job aggregation
Job filtering
AI-powered job-fit analysis
Groq LLM integration
Embedding infrastructure
Resume parsing infrastructure
Modern Streamlit UI
Light/dark themes
User profile interface
Application tracking infrastructure

The next development stages focus on completing the candidate profile, resume intelligence, semantic job matching, application tracking, and AI-assisted application workflow.

📄 License

This project is currently intended for educational and development purposes.

A formal open-source license can be added when the project is ready for public release.

AI Job Agent

Discover jobs. Understand your fit. Apply smarter.

Name		Name	Last commit message	Last commit date
app		app
backend		backend
config		config
frontend		frontend
tests		tests
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md

Latest commit

History

Folders and files

Repository files navigation