Free

Productionizing ML Systems without Fear nor Heroism

Actions and Detail Panel

Free

Event Information

Share this event

Date and time

Location

Location

Online event

Event description
Versioning data and models, testing data, and monitoring drift - Nastasia Saby

About this event

It’s time for machine learning systems to become real.

For that, we need stability in development, production, and maintenance. The distinctive characteristic of machine learning is that data and models are at least as important as the code. This is why we can take the best practices for code from software engineering and apply them to our data and models: versioning, monitoring, and tests.

Main topics

  • Versioning data manually with the concept of event sourcing or with a tool such as Delta Lake (quick demo)
  • Versioning models with a tool such as MLFlow (quick demo)
  • Test data manually or with a library such as GreatExpectations (quick demo)
  • Monitoring model drift and data drift

About the speaker:

Nastasia used to be a software backend developer. Now, She's a machine learning engineer specialised in productionizing models. Her role is to help others to productionize safely predictive systems.

She has a blog named Machine Learning in Real Life and a newsletter focused on data science in business, DataOps, and MLOps.

DataTalks.Club is the place to talk about data. Join our slack community!

Share with friends

Date and time

Location

Online event

Save This Event

Event Saved