Feature Request: Minimal AI model inference example for Kubernetes

[Ayush41]

A simple example demonstrating:

• A lightweight trained model (e.g., scikit-learn)
• A small inference API (FastAPI/Flask)
• Kubernetes manifests (Deployment, Service)
• Basic best practices (resource limits, probes)

would be valuable for users looking to understand how AI workloads are actually deployed on Kubernetes, without introducing heavy platforms or external dependencies.

This would complement existing and proposed AI examples (RAG, fine-tuning) by covering the core deployment pattern.