Applied build-and-operate course for containerized AI workloads; teaches deployment and operations, not model training or algorithm design. Deep cloud security is covered in AIINFRA 302; production inference serving and GPU orchestration continue in AIINFRA 200.
← All courses
// AIINFRA 101 · Semester 1
Applied Infrastructure & Containerization
Hands-on build-and-operate: Docker to Kubernetes to cloud, IaC, and CI/CD
A hands-on course in packaging, deploying, and operating AI workloads. Students progress from Docker images and GPU-enabled containers to deploying and scaling on Kubernetes, then move to the cloud: provisioning compute and storage, managing IAM and cost guardrails, using managed and serverless Kubernetes, infrastructure-as-code with Terraform, and CI/CD pipelines with GitHub Actions that automatically test, build, and deploy containerized model services.
download / canvas package
Canvas IMSCC Export
Canvas IMSCC Export
Download the complete Canvas import package for AIINFRA 101.
01 / outcomes
Outcomes
Course objectives
- Explain containerization and build optimized, secure Docker images for AI/Python applications.
- Compose and run multi-container and GPU-enabled AI stacks with the NVIDIA Container Toolkit.
- Deploy, scale, and manage AI workloads on Kubernetes (deployments, services, config, HPA, GPU-aware scheduling).
- Provision secure, cost-guarded cloud infrastructure (compute, GPU, storage, IAM) and manage it as code with Terraform.
- Build CI/CD pipelines with GitHub Actions that automatically test, build, deploy, and monitor containerized model services.
Student learning outcomes
- Build optimized, secure Docker images for AI/Python apps
- Run multi-container and GPU-enabled AI stacks (NVIDIA Container Toolkit)
- Deploy and scale AI workloads on Kubernetes (HPA, GPU-aware scheduling)
- Provision cost-guarded cloud infrastructure and manage it with Terraform
- Build CI/CD pipelines (GitHub Actions) for containerized model services
02 / schedule
16-week schedule
Wk 01
Why Containers for AI
VMs, images, and the Docker model; the kernel features that make containers work.
Wk 02
Docker Fundamentals
Running, inspecting, and managing containers; interactive vs. detached mode.
Wk 03
Building AI Images with Dockerfiles
Multi-stage builds and image security for smaller, safer AI images.
Wk 04
Docker Compose & Multi-Container Apps
From single containers to orchestrated multi-service AI stacks.
Wk 05
GPU Containers
Giving containers a GPU with the NVIDIA Container Toolkit.
Wk 06
Introduction to Kubernetes
Cluster architecture and kubectl fundamentals.
Wk 07
Kubernetes Workloads — Midterm
Deployments, services, and configuration; includes the course midterm.
Midterm · covers Wks 1–7Wk 08
Deploying an AI Inference App to Kubernetes
Serving a model on a cluster and reading GPU-bound vs. I/O-bound signals.
Wk 09
Scaling & GPU-Aware Scheduling
Horizontal Pod Autoscaler plus GPU-aware scheduling with the NVIDIA GPU Operator.
Wk 10
Cloud Foundations for AI
Providers, regions, free tiers, IAM least-privilege, and cost guardrails.
Wk 11
Cloud Compute & Storage
GPU instances, spot pricing, and object storage for AI data.
Wk 12
Managed & Serverless Kubernetes
Managed clusters and serverless compute for ML workloads in the cloud.
Wk 13
Infrastructure as Code with Terraform
Providers, resources, state, and provisioning cloud AI infrastructure as code.
Wk 14
CI/CD for ML with GitHub Actions
Automated testing, data/model quality gates, and image builds.
Wk 15
Automated Deployment, Monitoring & Cost
Blue-green/canary delivery, observability, and cloud cost optimization.
Wk 16
Final Project & Course Review
Students ship an end-to-end containerized, cloud-deployed service and review the course.
Capstone03 / Senate-ready controls
Course evidence and boundaries
These controls make the curriculum reviewable without changing the 10-course sequence.
A containerized AI service deployed and scaled on Kubernetes and the cloud through an automated CI/CD pipeline.
Image scanning/hardening and IAM least-privilege + cost-guardrail configuration.
Dockerized service, Kubernetes manifests/Helm or Terraform, a working CI/CD pipeline, and a monitoring/cost note.
04 / tools
Tools & frameworks
Containers
Orchestration
Cloud Platforms
Infrastructure as Code
CI/CD