Lilly Tech Systems
Learn Azure Kubernetes for AI
Master running AI and ML workloads on Azure Kubernetes Service. From AKS cluster setup and GPU node pools to KEDA-based auto-scaling and production model serving.
Your Learning Path
Follow these lessons in order, or jump to any topic that interests you.
1. Introduction
Why AKS for AI workloads? Benefits, architecture, and comparison with Azure ML managed compute.
2. AKS Setup
Create an AI-optimized AKS cluster with proper networking, identity, and monitoring configuration.
3. GPU Pools
Configure GPU node pools, install NVIDIA device plugins, and manage GPU scheduling with taints and tolerations.
4. KEDA Scaling
Event-driven auto-scaling with KEDA for inference workloads based on queue depth, HTTP traffic, and custom metrics.
5. Model Serving
Deploy models with Triton, TorchServe, and KServe on AKS with canary deployments and A/B testing.
6. Best Practices
Security, cost optimization, multi-tenancy, and operational excellence for AI workloads on AKS.
What You'll Learn
By the end of this course, you'll be able to:
Build AI Clusters
Set up production AKS clusters with GPU node pools, proper networking, and identity management.
Scale Intelligently
Use KEDA for event-driven scaling that matches inference demand automatically.
Serve Models
Deploy and manage ML models at scale with industry-standard serving frameworks.
Optimize Operations
Implement cost controls, security policies, and monitoring for production AI workloads.