AI Opportunity Finder Agent
Budget: ₹1,500 – ₹12,500 INR
I need an end-to-end AI agent that automatically scouts freelancing websites, general job boards, and specialised training platforms for roles or courses that involve artificial-intelligence work. The agent must:
• Crawl and scrape the relevant pages in real time or on a frequent schedule.
• Apply NLP or other classification techniques to decide whether a posting is truly AI-related, then tag it by sub-domain (e.g. vision, NLP, MLOps, prompt-engineering).
• Deliver concise, deduplicated listings to me through an in-app notification feed—no email or SMS required.
For the deployment side I’m open to Python (Scrapy, BeautifulSoup, Selenium), Node, or any stack you are comfortable with so long as it is containerised and can run unattended on a small cloud instance. A lightweight web interface or Electron desktop app for the notification feed is ideal; you can suggest an alternative if it achieves the same user experience.
Acceptance criteria
1. Agent successfully scrapes at least three live sources from each category—freelance, job board, and training.
2. Classification accuracy ≥ 90 % on a small validation set I will supply.
3. New items appear in the in-app feed within 15 minutes of being published online.
4. Deployment docs and a one-click setup (Docker-Compose or similar) are provided.
Please outline your proposed tech stack, any off-the-shelf APIs you expect to integrate, and the milestones you’d set to reach full production release.
• Crawl and scrape the relevant pages in real time or on a frequent schedule.
• Apply NLP or other classification techniques to decide whether a posting is truly AI-related, then tag it by sub-domain (e.g. vision, NLP, MLOps, prompt-engineering).
• Deliver concise, deduplicated listings to me through an in-app notification feed—no email or SMS required.
For the deployment side I’m open to Python (Scrapy, BeautifulSoup, Selenium), Node, or any stack you are comfortable with so long as it is containerised and can run unattended on a small cloud instance. A lightweight web interface or Electron desktop app for the notification feed is ideal; you can suggest an alternative if it achieves the same user experience.
Acceptance criteria
1. Agent successfully scrapes at least three live sources from each category—freelance, job board, and training.
2. Classification accuracy ≥ 90 % on a small validation set I will supply.
3. New items appear in the in-app feed within 15 minutes of being published online.
4. Deployment docs and a one-click setup (Docker-Compose or similar) are provided.
Please outline your proposed tech stack, any off-the-shelf APIs you expect to integrate, and the milestones you’d set to reach full production release.