Serving https://rohitesh-kumar-jain.github.io/Portfolio/

python -m venv venv

source venv/bin/activate

pip install -r requirements.txt

Hig Level Design

ai-platform-serv

• basically a cold-started serverless VM

So when someone hits site:

Cloudflare → Backend
           → backend is asleep   
           → cloud provider boots it
           → ~40–60 seconds
           → request finally reaches LLM

That’s why users see “no response” or timeouts sometimes.

Cloudflare Worker

Security perimeter

• Blocks bots and scrapers
• Shields your origin IP
• Hides your backend from the internet
• Can reject abusive requests before they reach your VM
• Can rate-limit per IP, country, ASN

Edge execution layer Handles CORS

• Handles OPTIONS preflight

AI firewall

• Your robots.txt and “no AI training” rules live here.
• Even if someone scrapes your site, they never reach your AI API.

Future Scope for Scale

Single backend -> Auto-scaling pool Cold starts -> Always-warm replicas 1 region -> Multi-region Manual limits -> Rate limits + quotas In-memory conversation -> Redis / Postgres