Model Online / On-Site
Duration
Modules 5 modules

Course Objective & Scope

When company data already lives in S3, Redshift or AWS Glue, managing ML servers manually outside AWS creates unnecessary security friction and idle-compute waste. This course teaches teams how to use Amazon SageMaker AI not merely as a hosted notebook, but as a managed, end-to-end enterprise MLOps platform.

Running ephemeral Processing and Managed Spot Training jobs via the SageMaker Python SDK, integrating custom Docker containers (BYOC), building automated DAGs with SageMaker Pipelines and Model Registry, choosing between Real-Time, Serverless, Asynchronous and Multi-Model Endpoints, and locking down VPC/IAM/KMS security are covered hands-on. Participants design a SageMaker MLOps architecture aligned with their AWS governance policies.

Who Should Attend?

  • Data scientists and MLOps engineers building and deploying models on AWS
  • Cloud architects and DevOps specialists migrating enterprise AI workloads to Amazon SageMaker
  • Platform teams looking to optimise SageMaker spend and network/security posture
  • Prerequisite: Python proficiency and basic familiarity with core AWS services (S3, IAM, ECR)

Key Takeaways

  • Run ephemeral preprocessing and Managed Spot Training jobs with the SageMaker Python SDK to eliminate idle GPU spend
  • Package custom training and inference code as Bring Your Own Container (BYOC) images in Amazon ECR
  • Automate preprocessing, training, evaluation gates and Model Registry registration with SageMaker Pipelines
  • Select and configure Real-Time, Serverless, Asynchronous or Batch Transform endpoints based on traffic patterns
  • Enforce enterprise security and observability with SageMaker Model Monitor, VPC-only mode, IAM roles and KMS encryption

Curriculum

SageMaker Architecture, Data Ingestion and BYOC Containers
  • SageMaker Studio, S3 input modes (FastFile / Pipe) and ephemeral compute provisioning
  • Pre-built Deep Learning Containers, Script Mode and the Bring Your Own Container (/opt/ml) contract with ECR
Distributed Training, Managed Spot Jobs and HPO
  • Multi-GPU/multi-node Training Jobs, automatic S3 checkpointing and resilient Managed Spot Training
  • SageMaker Automatic Model Tuning (AMT) and managed MLflow Tracking Server integration
MLOps Automation with SageMaker Pipelines and Model Registry
  • Authoring DAGs with ProcessingStep, TrainingStep, ClarifyCheckStep and ConditionStep
  • SageMaker Model Package Groups, approval statuses and EventBridge-triggered multi-account deployments
Flexible Inference Options and Cost Optimisation
  • Choosing between Real-Time, Serverless, Asynchronous Inference and Batch Transform
  • Packing dozens of models onto shared instances with Multi-Model Endpoints (MME) and Inference Components
Model Monitor, Enterprise Security (VPC/IAM) and FinOps
  • Enabling endpoint Data Capture, scheduling SageMaker Model Monitor baselines and wiring CloudWatch alarms
  • VPC-only isolation via PrivateLink, KMS encryption, least-privilege IAM Execution Roles and cost allocation tags

Other Courses in This Category

  • Production Model Monitoring and ML Observability

    System and model telemetry with OpenTelemetry, Prometheus and Grafana, prediction logging, delayed ground-truth joining, live performance tracking and incident response, for MLOps and SRE teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Data and Concept Drift Detection and Continuous Adaptation

    The maths of covariate, label and concept drift; PSI, KS, Wasserstein and MMD tests; label-free performance estimation with Evidently/NannyML and retraining strategies, for data scientists and MLOps teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Feature Store Architectures and Feature Engineering Platforms

    Feature Store architectures preventing training-serving skew and time leakage: offline (Parquet/Snowflake/BigQuery) and online (Redis) stores with Feast, point-in-time joins and streaming features, for data and MLOps teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Responsible and Safe ML Deployment

    Pre-deployment fairness (Fairlearn), explainability (SHAP), probability calibration, uncertainty quantification (Conformal Prediction), guardrails, progressive rollouts and kill-switches, for ML and risk teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Enterprise AI Platform Architecture and Platform Engineering

    Designing an enterprise AI platform unifying classical ML and GenAI/LLMs: compute/GPU orchestration, data and model layers, AI Gateways, multi-tenant isolation and reference architectures, for architects and platform leads.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Self-Service AI Platforms and Internal Developer Experience (Golden Paths)

    Designing an Internal Developer Platform (IDP) for data scientists and AI engineers: Backstage templates, on-demand notebook/GPU workspaces, Golden Path pipelines and self-service deployment, for platform teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View

Get in touch