Get in Touch
 Duration 14 hours

Course Outline

Core Principles of MLOps on Kubernetes

  • Essential concepts in MLOps
  • Contrasting MLOps with traditional DevOps
  • Primary challenges in managing the ML lifecycle

Containerization of ML Workloads

  • Encapsulating models and training scripts
  • Enhancing container images for ML efficiency
  • Handling dependencies to ensure reproducibility

CI/CD for Machine Learning

  • Organizing ML repositories for automated processes
  • Incorporating testing and validation stages
  • Initiating pipelines for retraining and model updates

GitOps for Model Release

  • Foundations and workflows of GitOps
  • Utilizing Argo CD for model deployment
  • Version management for models and configurations

Orchestrating Pipelines on Kubernetes

  • Constructing pipelines using Tekton
  • Oversight of multi-step ML workflows
  • Scheduling tasks and managing resources

Monitoring, Logging, and Rollback Protocols

  • Monitoring data drift and model efficacy
  • Integrating alert systems and observability tools
  • Implementing rollback and failover methods

Automated Retraining and Continuous Enhancement

  • Creating effective feedback loops
  • Automating scheduled retraining processes
  • Integrating MLflow for tracking and experiment oversight

Advanced MLOps Architectures

  • Deployment models for multi-cluster and hybrid-cloud environments
  • Enabling team scaling via shared infrastructure
  • Addressing security and compliance requirements

Conclusion and Future Directions

Requirements

  • Familiarity with core Kubernetes concepts
  • Practical experience with machine learning workflows
  • Proficiency in Git-based development practices

Target Audience

  • Machine Learning Engineers
  • DevOps Engineers
  • ML Platform Teams

Number of participants


Price per participant

Testimonials (3)

Upcoming Courses

Related Categories