Microsoft MVP | Solutions Architect and Engineer
I build and share practical demos for Azure AI, Azure Kubernetes Service (AKS), Kubernetes-based AI workloads, and Microsoft Foundry.
My projects are designed for developers, platform engineers, and architects who want to move from product concepts to working implementations. I use these repositories in technical articles, YouTube videos, and community presentations.
- Running open-source language models on AKS with KAITO and vLLM
- Operating GPU inference workloads with autoscaling, observability, and cost controls
- Building and evaluating agents with Microsoft Foundry and Microsoft Agent Framework
- Connecting cloud-native tools and agents through MCP and A2A
- Applying responsible AI practices through evaluation, guardrails, and red teaming
| Repository | What you will learn |
|---|---|
| aks-demos | Deploy and operate AI workloads on AKS with KAITO, RAGEngine, ingress, and Azure Key Vault |
| azure-ai-apps | Build RAG applications, AI agents, evaluation workflows, and chat experiences on Azure |
| multi-agentic-loop-lab | Coordinate specialized GitHub Copilot agents through a traceable software-development loop |
| az-terraform-cicd | Explore AKS infrastructure patterns using Terraform, Azure Verified Modules, and CI/CD |
Each flagship project is intended to become a complete learning package:
research -> GitHub demo -> animated slides -> article -> meetup talk -> video -> short clips
My next series covers:
- KAITO and vLLM inference on AKS
- Hosted agents with Microsoft Foundry
- Production inference operations on Kubernetes
- AKS-hosted MCP tools for Foundry agents
- Eval-driven agent delivery
- Responsible AI and red-team testing
- Read: Roy Kim on Azure and AI
- Watch: youtube.com/roykimyyz
- Discuss: x.com/roykimyyz
- Explore: All public repositories
Questions, corrections, and documentation improvements are welcome in the relevant repository.