Taras Svirskyi

05/13/2022, 3:50 PM
Hi everyone! My name is Taras, and Iโ€™m a data scientist. Iโ€™ve decided to take a closer look at Prefect for managing ML modelโ€™s training and backtesting pipelines. I used Luigi and DVC in the past and am keen to explore python-native dynamic orchestration now ๐Ÿ™‚ Iโ€™m also interested in running flows in the cloud, so I will probably ask questions about infrastructure.
๐Ÿ‘‹๐Ÿฝ 1
๐Ÿ‘‹ 14

Kevin Kho

05/13/2022, 3:50 PM
Hi @Taras Svirskyi! Welcome to Prefect!
๐Ÿ‘‹ 1

Henning Holgersen

05/13/2022, 4:12 PM
Hi, welcome!
๐Ÿ‘‹ 1

Chris Reuter

05/13/2022, 4:27 PM
Welcome to the community @Taras Svirskyi! DVC is interesting, were you using it in combination with Luigi or standalone?
๐Ÿ‘‹ 1


05/13/2022, 4:42 PM
Hey, welcome @Taras Svirskyi! ๐Ÿ‘‹ Joined us just in time for Meme Friday over in #random ๐Ÿ˜€
๐ŸŽ‰ 2

Taras Svirskyi

05/13/2022, 4:48 PM
Iโ€™m using DVC standalone. Luigi was my first attempt to manage ML code but I end up writing too many boilerplate in attempt to support dynamic pipelines. The DVC provides nice abstractions for metrics/parameters/steps of the pipelines + cache. Yet it becomes quite hard to manage all steps as pipelines grow cause itโ€™s yaml+python.
๐Ÿ‘ 2

Anna Geller

05/13/2022, 4:52 PM
Welcome aboard, @Taras Svirskyi! Great to have you with us!
๐Ÿ‘‹ 1