CloudSketch AI Logo
FeaturesTemplatesPricingEnterpriseAboutContactLog in
πŸŒ™β˜€οΈ
Log inStart free
Home/Templates/AI & Machine Learning

🧠 AI & Machine Learning · Deployment · Pro

MLOps Model Training Pipeline

Where each part of an MLOps setup runs: feature store, training jobs on GPU nodes, experiment tracking, model registry and automated deployment to serving.

More AI & Machine Learning templates

Deployment diagrams are part of Pro. Anyone can view this one; generating and editing it needs Pro.

Drawing diagram…

What this diagram shows

  • Training runs on a separate GPU node pool
  • Every model version is registered with its metrics
  • Only approved models are promoted to production

Prompt used

Deployment of an MLOps platform on Google Cloud: data lands in BigQuery and a Feast feature store. Kubeflow Pipelines on GKE runs training jobs on a GPU node pool, logs runs to MLflow, and registers models in the MLflow model registry. After approval, Argo CD deploys the model to a KServe serving cluster behind a load balancer. Models and artifacts are stored in Cloud Storage.

Mermaid code
flowchart TB
  DS[Data Scientists] -->|"Pipeline runs, HTTPS"| KF
  subgraph GCP[Google Cloud]
    BQ[(BigQuery)]
    FS[(Feast Feature Store)]
    GCS[(Cloud Storage - artifacts)]
    subgraph TRAIN[GKE Training Cluster]
      KF[Kubeflow Pipelines]
      subgraph GPU[GPU Node Pool]
        JOB[Training Jobs]
      end
      MLF[MLflow Tracking and Registry]
    end
    subgraph SERVE[GKE Serving Cluster]
      ARGO[Argo CD]
      KS[KServe Model Pods]
    end
    LB[Load Balancer]
  end
  BQ -->|"Feature tables"| FS
  KF -->|"Launch training jobs"| JOB
  FS -->|"Training features"| JOB
  JOB -->|"Metrics and model versions"| MLF
  JOB -->|"Model artifacts"| GCS
  MLF -->|Approved version| ARGO
  ARGO -->|"Deploy InferenceService"| KS
  GCS -->|"Load model artifacts"| KS
  LB -->|"HTTPS predictions"| KS
  APPS[Client Apps] -->|"HTTPS :443"| LB

Inside this diagram

A deployment diagram with 11 components in 4 environments and 11 connections.

Environments and what runs in them

  • Google Cloud: BigQuery, Feast Feature Store, Cloud Storage - artifacts and Load Balancer
  • GKE Training Cluster: MLflow Tracking and Registry
  • GPU Node Pool: Training Jobs
  • GKE Serving Cluster: Argo CD and KServe Model Pods

Outside these environments

  • Data Scientists
  • Kubeflow Pipelines
  • Client Apps

Data stores

  • BigQuery
  • Feast Feature Store
  • Cloud Storage - artifacts

Connections

  • Data Scientists β†’ Kubeflow Pipelines: Pipeline runs, HTTPS
  • BigQuery β†’ Feast Feature Store: Feature tables
  • Kubeflow Pipelines β†’ Training Jobs: Launch training jobs
  • Feast Feature Store β†’ Training Jobs: Training features
  • Training Jobs β†’ MLflow Tracking and Registry: Metrics and model versions
  • Training Jobs β†’ Cloud Storage - artifacts: Model artifacts
  • MLflow Tracking and Registry β†’ Argo CD: Approved version
  • Argo CD β†’ KServe Model Pods: Deploy InferenceService
  • Cloud Storage - artifacts β†’ KServe Model Pods: Load model artifacts
  • Load Balancer β†’ KServe Model Pods: HTTPS predictions
  • Client Apps β†’ Load Balancer: HTTPS :443

Make your own diagram like this

Describe your own system in plain English and the AI cloud architecture diagram generator draws it in seconds, in the same style as this template.

Related templates

C4 ArchitectureProAI & Machine Learning

RAG Chatbot over Company Documents

A chatbot that answers staff questions from company documents: documents are split and indexed in a vector database, and each question pulls the most relevant passages before the language model writes an answer.

SequenceAI & Machine Learning

RAG Question Answering Sequence

What happens when a user asks the document chatbot a question: permission check, embedding, vector search, prompt building and a cited answer.

Data FlowAI & Machine Learning

Document Ingestion for Vector Search

How documents become searchable passages for an AI assistant: extraction, cleaning, chunking, embedding and indexing, with changed files re-processed automatically.

FlowchartAI & Machine Learning

AI Agent Tool-Use Workflow

How an AI agent completes a task by planning, calling tools, checking results and asking a person to approve risky actions before it finishes.

State MachineAI & Machine Learning

ML Model Lifecycle

The states a machine learning model moves through, from experiment to production to retirement, including rollback when live performance drops.

C4 ArchitectureProAI & Machine Learning

LLM Gateway with Cost and Safety Controls

One gateway that every app in a company uses to reach language models: it applies budgets, rate limits, content filters and caching, and records usage per team.