Deployment of ML models into production

How it went

We focused on what follows after training the model – that is, on the path from the experiment to production deployment and real operation.

In this workshop, the focus shifted from modeling to deployment. The participants went through the main technical steps that separate a promising experiment from a stable production solution: model packaging, inference pipeline, monitoring, scaling and working with real-time data. Also discussed were operational trade-offs, performance optimization, and how to keep models healthy after deployment. The workshop helped place machine learning in the broader framework of MLOps and practical operation.

Event resources

• Photo Gallery

Event speakers

• Dmytro Molokoiedov