AI Job Acquisition Engine

A config-driven Telegram job intelligence system for AI/backend engineers.

AI Job Acquisition Engine monitors Telegram channels and RSS feeds, scores opportunities against your profile, and generates response drafts you can manually review before sending.

Features

• Telegram job listener (Telethon)
• RSS ingestion pipeline
• Editable YAML skill matcher
• Relevance scoring engine with configurable weights
• SQLite tracking and status management
• Proposal draft generator (template-based)
• Manual-approval workflow (no auto-spam)
• CLI dashboard + FastAPI review dashboard
• Owner-only Telegram UI bot (ui_bot.py) with approve/skip/applied actions

Architecture

Telegram Channels + RSS Feeds
            |
            v
   Listener Layer (Telethon / RSS Poller)
            |
            v
 Matcher (YAML-configurable scoring rules)
            |
            v
      SQLite Storage (jobs.db)
            |
            v
 Notification + Draft Generator
            |
            v
   Review Interfaces (CLI / FastAPI)

Setup

1. Install dependencies

pip install -r requirements.txt

2. Configure Telegram credentials

Create your credentials file and add:

• api_id
• api_hash
• bot_token (from @BotFather)
• owner_user_id (your Telegram user id; UI bot is owner-locked)

3. Validate Telegram session manually

from telethon.sync import TelegramClient

api_id = 123456
api_hash = "your_api_hash"

client = TelegramClient("session", api_id, api_hash)
client.start()
print(client.get_me())

If this succeeds, auth/session setup is correct.

4. Initialize local database

python3 init_db.py

5. Run collectors

python3 bot.py
python3 rss_poller.py

Run the owner UI bot in a second terminal:

python3 ui_bot.py

Commands:

• /start
• /inbox
• /strong
• /stats

6. Review matched jobs

CLI dashboard:

python3 dashboard_cli.py --min 14 --status pending --limit 30

Web dashboard:

uvicorn dashboard_web:app --reload

Open http://127.0.0.1:8000.

Example Scoring Config

thresholds:
  notify: 14
  strong: 22

weights:
  must_have: 6
  nice_to_have: 3
  role_match: 4
  negative: -8

keywords:
  must_have:
    - python
    - fastapi
    - postgresql
  nice_to_have:
    - docker
    - aws
    - llm

negative_keywords:
  - unpaid
  - commission-only

Notification Screenshot

Roadmap

• Telegram listener
• Config-driven skill matching
• Draft generation (template-based)
• Web dashboard (FastAPI)
• Multi-source support (RSS)
• Analytics + conversion tracking
• Team review workflow
• Hosted SaaS foundation (Postgres + auth)

Safety

• No auto-apply and no automated unsolicited outreach
• Human review required before sending any response
• Keep credential/session files private (*.session, API keys)

Positioning

Built as a practical foundation for an AI job intelligence product: local-first today, SaaS-ready architecture tomorrow.