Скачать или смотреть Designing a Model Registry with Legacy Systems Using DVC and GTO with Francecso Calcavecchia

Francesco has contributed a ton to GTO ([https://iterative.ai/gto](https://iterative.ai/gto)), one of the components of our [MLEM.ai](http://mlem.ai/) tool to help create a model registry and then easily package and deploy your models for production, all while tracking everything with Git.

In his talk, Francesco illustrates how he used DVC and GTO to build a combined model & experiment registry. The resulting design is highly atypical but was the natural consequence of the legacy setup, and it may apply to other ML projects with unconventional MLOps practices. In the last part of the talk, he sketches how such a setting paves the way for a tracked event-driven architecture.

To learn more about Iterative's open-source and SaaS tools please visit:
🧑🏽‍💻 Our online course: https://learn.iterative.ai
✍🏼 Our docs:[https://dvc.org/doc
https://cml.dev/doc
https://mlem.ai/doc
https://studio.iterative.ai

Join our Discord server:   / discord  

Resources shared in the chat:

Francesco on LinkedIn: www.linkedin.com/in/francescocalcavecchia/

Naughty or nice blog post using MLEM:   / i-trained-a-model-to-tell-if-you-were-naug...