About this Event
Learn how to build scalable, reproducible ML workflows using industry-standard open source tools. Whether you're working with traditional models or large language models - this hands-on workshop covers real code, real results, real deployments. (PowerPoint slides kept to an absolute minimum!)
Key Benefits
- Platform Independent: Use open source tools only, avoid vendor lock-in
- Production Ready: Learn proven workflows that scale from your machine to enterprise clusters
- Hands-on Learning: Build and deploy scalable and reproducible ML pipelines yourself, not just theory
- Expert Guidance: Learn from prokube's MLOps engineers with years of experience
What You'll Learn
By the end of this workshop, you'll be able to:
- Build scalable ML training pipelines with Kubeflow Pipelines for reproducible data preprocessing and model training
- Track experiments systematically with MLflow for ML experiment tracking and model management
- Deploy trained models to production using KServe for high-performance, higly available and scalable model serving
- Set up model registry workflows and configure deployment automation
Who Should Attend
Perfect for:
- Python Developers transitioning into ML engineering
- Data Scientists wanting to productionize their models
- ML Engineers seeking to master industry-standard tools
- Team Leads planning MLOps implementations
Prerequisites:
- Python experience (intermediate level)
- Basic understanding of containerization (Docker concepts)
- ML fundamentals (model training, evaluation basics)
- Laptop with webbrowser installed
Not required:
- Kubernetes expertise (we'll cover the basics)
- Prior MLOps experience
- Specific ML framework knowledge
- Local development environment
*Our confidence guarantee - we're sure you'll find value that we'll refund your fee if you didn't feel you got your money's worth - no questions asked (only if you attendet the complete workshop).
Questions?
Contact us at [email protected]
Agenda
Welcome & MLOps Fundamentals (30 min)
Info: 1) Introductions and networking; 2) Why MLOps matters: From notebook chaos to production success; 3) Overview of the open source MLOps ecosystem
Technical Foundation (30 min)
Info: 1) Container & Kubernetes essentials for ML; 2) Introduction to Kubeflow architecture; 3) Setting up your development environment
Kubeflow Pipelines Deep Dive (2 hours)
Info: 1) Theory (30 min): Pipeline concepts, components, and workflows; 2) Hands-on (90 min): Launch your first (Kubeflow) notebook environment, build lightweight pipeline components, create your first ML pipeline, execute, debug and monitor pipeline runs
Lunch Break (12:00-14:00)
Info: Networking lunch with fellow participants and prokube team
MLflow Integration (45 min)
Info: 1) Experiment Tracking: Organize and compare model experiments; 2) Model Registry: Manage model versions and deployments; 3) Hands-on: Integration with Kubeflow workflows
Model Serving with KServe (75 min)
Info: 1) Theory (30 min): Production serving requirements and KServe architecture; 2) Hands-on (45 min): Deploy and test your trained models
Advanced Topics & Next Steps (45 min)
Info: 1) Containerized components for complex workflows; 2) Hyperparameter tuning with Katib; 3) Multi-GPU multi-node training with Training Operator; 4) Custom transformers and serving logic; 5) Building your MLOps roadmap
Wrap-up & Q&A (15 min)
Info: 1) Key takeaways and best practices; 2) Resources for continued learning; 3) prokube platform demonstration
Event Venue & Nearby Stays
WeWork - Axel-Springer-Platz - Private Büroflächen und Coworking, Axel-Springer-Platz 3, Hamburg, Germany
EUR 300.00
