Add prompt engineering tutorial blueprint with Ministral-8B on vLLM

[tugglep]

What does this PR do?

Adds a new prompt engineering tutorial blueprint that deploys Ministral-8B on vLLM with an OpenAI-compatible API. Includes 10 Jupyter notebooks covering prompt engineering fundamentals from basic prompt structure to reducing hallucinations.

Motivation

Created for the EKS Startups Program to provide hands-on prompt engineering education using self-hosted LLMs on EKS. This enables developers to learn prompt engineering techniques using infrastructure they control, with no external API dependencies.

More

• Yes, I have tested the PR using my local account setup (Provide any test evidence report under Additional Notes)
• Mandatory for new blueprints. Yes, I have added a example to support my blueprint PR
• Mandatory for new blueprints. Yes, I have updated the website/docs or website/blog section for this feature
• Yes, I ran pre-commit run -a with this PR

Additional Notes

Test Evidence:

• Deployed on EKS cluster (prompt-eng) with Karpenter auto-provisioning g6e.2xlarge GPU node
• Verified vLLM endpoint via curl http://localhost:8000/v1/models
• Successfully ran notebooks 00_setup, 01_basic_prompt_structure, and 02_role_and_purpose

Blueprint contents:

• ministral-8b-vllm.yaml - K8s manifest for vLLM deployment
• README.md - Quick start guide with architecture, prerequisites, and troubleshooting
• 10 Jupyter notebooks (00-09) covering prompt engineering techniques

Note: Website docs update pending - happy to add if maintainers confirm desired location.

[omrishiv]

Hi @tugglep, thank you for opening this PR, this looks very thorough! Rather than add another deployment for vLLM, can you take a look at awslabs/ai-on-eks-charts. We have a simple way to deploy many inference servers using this helm chart.

Additionally, rather than a set of notebooks, does it make sense to convert this to Guidance. As it is, reading/executing these notebooks is challenging without having Jupyter deployed.