This hands-on, build-and-operate course teaches students to package, deploy, scale, and automate AI and Python workloads across the full modern infrastructure stack. Students begin by containerizing applications with Docker — building optimized, secure images and composing multi-container, GPU-enabled stacks with the NVIDIA Container Toolkit — then orchestrate those workloads on Kubernetes with deployments, services, configuration, autoscaling, and GPU-aware scheduling. From there the course moves into the cloud: provisioning cost-guarded compute, GPU instances, object storage, and IAM across the major providers, running managed and serverless Kubernetes, and managing all of it as code with Terraform. The course closes by wiring everything together with CI/CD pipelines in GitHub Actions that automatically test, build, deploy, monitor, and cost-optimize containerized model services. By the end, students can take an AI application from a Dockerfile on their laptop to an automated, observable, cost-aware deployment in the cloud.
54 contact hours (3-unit equivalent), delivered over a 16-week term at approximately 3.4 hours per week; non-credit, CDCP Certificate of Completion pathway.
AIINFRA 100, or equivalent working knowledge of Python, Git, and the command line.
By the end of this course, students will be able to:
| Week | Topic |
|---|---|
| 01 | Why Containers for AI: VMs, Images, and the Docker Model |
| 02 | Docker Fundamentals: Running & Managing Containers |
| 03 | Building AI Images with Dockerfiles: Multi-Stage & Image Security |
| 04 | Docker Compose & Multi-Container AI Applications |
| 05 | GPU Containers: the NVIDIA Container Toolkit |
| 06 | Introduction to Kubernetes: Architecture & kubectl |
| 07 | Kubernetes Workloads: Deployments, Services, Config (Midterm) |
| 08 | Deploying an AI Inference App to Kubernetes |
| 09 | Scaling & GPU-Aware Scheduling: HPA and the NVIDIA GPU Operator |
| 10 | Cloud Foundations for AI: Providers, IAM & Cost Guardrails |
| 11 | Cloud Compute & Storage: GPU Instances, Spot, Object Storage |
| 12 | Managed & Serverless Kubernetes in the Cloud |
| 13 | Infrastructure as Code with Terraform |
| 14 | CI/CD for ML with GitHub Actions |
| 15 | Automated Model Deployment, Monitoring & Cost Optimization (Final) |
| Component | Weight |
|---|---|
| Labs | 40% |
| Discussions | 10% |
| Weekly Quizzes | 15% |
| Midterm | 15% |
| Final | 20% |
This course is graded Credit / No-Credit. A cumulative score of 70% or higher is required to earn Credit and progress toward the Certificate of Completion.
Academic integrity: Students are expected to submit their own work. Collaboration on concepts and troubleshooting is encouraged, but each student must independently configure, run, and submit their own labs, containers, cloud deployments, and pipelines. Submitting another student's work — code, configuration, screenshots, or quiz answers — as your own is a violation of the LACCD Student Conduct Code and may result in a zero on the assignment and referral to the college's student conduct process. Late work: Assignments are due as posted in Canvas. Late labs and quizzes are accepted for up to one week after the due date for reduced credit, at the instructor's discretion. The Midterm and Final follow stricter deadlines due to their cumulative nature; extensions require instructor approval in advance. Students facing extenuating circumstances should contact the instructor as soon as possible. Responsible use of AI: Students may use AI tools (e.g., Claude, ChatGPT, Copilot) to support learning — explaining errors, brainstorming approaches, and accelerating configuration — consistent with real-world industry practice. Students must be able to explain and reproduce any code, configuration, or pipeline they submit in their own words, and must not submit AI output they do not understand. Instructors may ask students to walk through their work verbally. Accessibility: LACCD is committed to ensuring equal access. Students with disabilities who may need accommodations should contact the campus Disabled Students Programs and Services (DSPS) office as early in the term as possible. Course materials are provided in accessible formats upon request.
All tools used in this course are free or available on a free tier — no purchases are required. Students should be able to install and run:
kubectl plus a local cluster (kind or minikube), Deployments/Services/Ingress, Horizontal Pod Autoscaler, the NVIDIA GPU Operatoract as a local runner), container image build/push