MLOps & Deployment Lab

Overview

The MLOps Lab bridges the gap between a trained model on your laptop and a reliable ML system in production. You get a real GKE Kubernetes cluster, Vertex AI Pipelines for workflow orchestration, Cloud Build for CI/CD, and a full monitoring stack (Prometheus + Grafana). This is the exact infrastructure pattern used by ML platform teams at production companies. You'll build automated training pipelines, deploy models with canary releases, set up data drift monitoring, and implement the full ML lifecycle from commit to production.

What You'll Do in This Lab

Build Vertex AI Pipelines (Kubeflow) for automated training and evaluation
Deploy models to GKE with KServe — auto-scaling, health checks, and versioning
Set up MLflow for experiment tracking, model registry, and artifact storage
Create Cloud Build CI/CD: push code → run tests → retrain → validate → deploy
Implement A/B testing and canary deployments with traffic splitting
Build Grafana dashboards for model latency, throughput, and data drift monitoring

Lab Workflow

1

Launch

Start the lab. A GKE cluster, Vertex AI Pipelines environment, MLflow server, Cloud Build triggers, and monitoring stack are provisioned.

2

Pipeline

Define a Vertex AI Pipeline that downloads data, trains a model, evaluates it, and conditionally deploys if quality thresholds are met.

3

CI/CD

Connect a Cloud Build trigger to your Git repository. Push code changes and watch automated training and deployment kick off.

4

Deploy

Deploy your model to GKE with KServe. Configure auto-scaling rules, health probes, and traffic routing.

5

Monitor

Open Grafana dashboards. Send test traffic and observe latency, throughput, and prediction distribution metrics in real-time.

6

Iterate

Push a model update. Watch the canary deployment roll out, monitor for regressions, and roll back or promote.

Hardware & Environment

GKE Cluster	3x e2-standard-4 nodes (4 vCPU, 16 GB RAM each)
Pipeline Engine	Vertex AI Pipelines (managed Kubeflow Pipelines)
CI/CD	Cloud Build with Artifact Registry for Docker images
Experiment Tracking	MLflow on GKE with Cloud SQL backend
Monitoring	Prometheus + Grafana on GKE
Session Length	Persistent environment — deployments stay running between sessions

Frequently asked questions about this lab

What is the MLOps & Deployment Lab? +

Production ML infrastructure environment. Students build CI/CD pipelines for ML, deploy models to Kubernetes, set up monitoring, and implement A/B testing and canary deployments.

Which courses use this lab? +

This lab is included in: MLOps & AI Infrastructure.

What hardware does this lab run on? +

Vertex AI Pipelines + GKE + Cloud Build. GKE Cluster: 3x e2-standard-4 nodes (4 vCPU, 16 GB RAM each); Pipeline Engine: Vertex AI Pipelines (managed Kubeflow Pipelines); CI/CD: Cloud Build with Artifact Registry for Docker images; Experiment Tracking: MLflow on GKE with Cloud SQL backend.

What software comes pre-installed? +

Comes pre-loaded with Vertex AI Pipelines (Kubeflow), MLflow on GKE, Seldon Core / KServe, Cloud Build + Artifact Registry, Prometheus + Grafana, Terraform. No local installs or dependency setup required — open your browser and start working.

Can I bring my own datasets and code into this lab? +

Yes. Datasets can be uploaded directly or synced from Google Cloud Storage. Notebooks and source files have built-in Git integration so you can push work to your own GitHub or GitLab repos.

Do I need to enroll in a course to use this lab? +

Yes. Lab environments are provisioned per-student as part of an AI Labs course enrollment. Browse the courses linked above to find programs that include this lab.

Related labs

Other AI Labs environments students typically use alongside this one.

Ready to Try This Lab?

Enroll in a course that uses this lab, or visit our Houston center for a hands-on demo.

Browse Courses View All Labs