How georgeknapendatech/linkedin_auto_jobs_applier_with_ai_fast works
Project type and purpose
• LinkedIn_AIHawk is an open‐source, command-line automation tool (a “bot”) that accelerates and personalizes your LinkedIn job search and application process. It leverages browser automation (via Selenium/ChromeDriver), OpenAI’s GPT API, and dynamic resume generation to auto-fill forms, tailor resumes, and submit “Easy Apply” applications at scale.
File structure
Top-level files:
• main.py – entry point and CLI argument parsing
• gpt.py – wrappers around OpenAI API for AI-driven text generation
• linkedIn_authenticator.py – logs into LinkedIn via Selenium
• linkedIn_bot_facade.py – high-level façade coordinating authentication, search, application
• linkedIn_easy_applier.py – implements “Easy Apply” form interactions
• linkedIn_job_manager.py – orchestrates job search, filtering, and queuing
• resume.py – builds and renders tailored resume PDFs using template scripts
• utils.py, strings.py – shared helper functions and constant strings
• data_folder/ – user’s secrets.yaml, config.yaml, plain_text_resume.yaml (runtime inputs)
• resume_template/ – Node.js scripts (casual_markdown.js, reorganizeHeader.js) and CSS for HTML→PDF
• documentation.md, README.md, requirements.txt, LICENSE
Patterns & organization:
• Facade pattern in linkedIn_bot_facade.py
• Separation of concerns: authentication, job management, application, resume generation, GPT integration
• External configs (YAML) isolate sensitive and user-specific data
• Template directory uses a mini Node.js pipeline for HTML/CSS→PDF transform
Insights from the README
• Core flows: configure your bot via YAML files; run main.py; bot logs into LinkedIn; scans and filters jobs; generates AI-personlized text; generates or uses a PDF resume; auto-submits applications.
• Emphasis on “data_folder_example” to illustrate correct inputs.
• External dependencies: Python packages (Selenium, OpenAI), Chrome & ChromeDriver, Node.js for resume templates.
To capture the architecture, break the system into these main components:
A. User Interface (CLI)
• Accepts flags (e.g., --resume)
• Reads and validates config/data folder
B. Configuration & Data Store
• YAML file loader for secrets.yaml, config.yaml, plain_text_resume.yaml
• Example data folder for reference
C. Bot Controller / Facade
• linkedIn_bot_facade.py orchestrates end-to-end workflow
D. Authentication Service
• linkedIn_authenticator.py uses Selenium/ChromeDriver to log in
E. Job Manager
• linkedIn_job_manager.py handles search criteria, continuous scanning, blacklisting
F. Application Service
• linkedIn_easy_applier.py executes “Easy Apply” flows, form autofill
G. GPT Service
• gpt.py communicates with OpenAI API for question responses and personalization
H. Resume Generation
• resume.py composes HTML via plain_text_resume.yaml
• Calls Node.js scripts (casual_markdown.js, reorganizeHeader.js) + CSS → outputs PDF
I. Browser Automation
• Selenium + ChromeDriver as the runtime engine for all LinkedIn interactions
J. External Services
• LinkedIn Web UI
• OpenAI API
Relationships & data flows:
1. CLI → loads Configuration & Data
2. CLI → Bot Controller
3. Bot Controller → Authentication Service → LinkedIn
4. Bot Controller → Job Manager → LinkedIn (search pages)
5. Job Manager → Application Service → Browser Automation → LinkedIn (forms)
6. Application Service → GPT Service for dynamic text
7. Bot Controller → Resume Generation → Node.js template pipeline → PDF
8. Application Service → attaches PDF + submits
Architectural patterns & principles:
• Facade (linkedIn_bot_facade.py) to hide complexity
• Modular services: each major concern in its own module
• Configuration-driven behavior via YAML (separation of code & data)
• External template pipeline (Node.js) for resume styling
Key elements to capture
• Clear component labels and boundaries
• Data/config sources shown as separate nodes