Course Outline
Foundations of MLOps on Kubernetes
- Core concepts of MLOps.
- Differences between MLOps and traditional DevOps.
- Key challenges in managing the ML lifecycle.
Containerizing ML Workloads
- Packaging models and training code.
- Optimizing container images for ML purposes.
- Managing dependencies and ensuring reproducibility.
CI/CD for Machine Learning
- Structuring ML repositories to support automation.
- Integrating testing and validation steps.
- Triggering pipelines for retraining and updates.
GitOps for Model Deployment
- GitOps principles and workflows.
- Utilizing Argo CD for model deployment.
- Maintaining version control for models and configurations.
Pipeline Orchestration on Kubernetes
- Building pipelines using Tekton.
- Managing multi-step ML workflows.
- Handling scheduling and resource management.
Monitoring, Logging, and Rollback Strategies
- Tracking data drift and model performance.
- Integrating alerting and observability tools.
- Implementing rollback and failover approaches.
Automated Retraining and Continuous Improvement
- Designing feedback loops.
- Automating scheduled retraining processes.
- Integrating MLflow for tracking and experiment management.
Advanced MLOps Architectures
- Multi-cluster and hybrid-cloud deployment models.
- Scaling teams through shared infrastructure.
- Addressing security and compliance considerations.
Summary and Next Steps
Requirements
- A solid understanding of Kubernetes fundamentals.
- Prior experience with machine learning workflows.
- Knowledge of Git-based development practices.
Audience
- ML engineers.
- DevOps engineers.
- ML platform teams.
Testimonials (3)
The knowledge and the patience from the trainer to answer to our questions.
Calin Avram - REGNOLOGY ROMANIA S.R.L.
Course - Deploying Kubernetes Applications with Helm
the ML ecosystem not only MLFlow but Optuna, hyperops, docker , docker-compose
Guillaume GAUTIER - OLEA MEDICAL
Course - MLflow
I enjoyed participating in the Kubeflow training, which was held remotely. This training allowed me to consolidate my knowledge for AWS services, K8s, all the devOps tools around Kubeflow which are the necessary bases to properly tackle the subject. I wanted to thank Malawski Marcin for his patience and professionalism for training and advice on best practices. Malawski approaches the subject from different angles, different deployment tools Ansible, EKS kubectl, Terraform. Now I am definitely convinced that I am going into the right field of application.
