Hyperight

Our model stopped working! Now what? A credit card fraud multi-persona MLOPs story – Iulia Feroli, Dataiku

Premium content

Login or register to unlock the content

Session Outline

There are many approaches and strategies for MLOPs in data science departments, each with their own focus, benefits, or pain points. From standardizing practices and way of work, to automating the boring tasks of your job, to extensive performance monitoring, to governance and control: MLOPs can take many forms, especially depending on which roles you talk to. We believe a successful MLOPs implementation includes everybody. In this demo we will walk through a crisis scenario and showcase how different roles approach drift detection, model iteration, and governance policies; all in the same platform, to ensure continuous business results and performance.

Key Takeaways

– Governance is a key aspect of MLOPs; we highlight some aspects of compliance you should watch out for / include in your strategy

– Model and data drift monitoring strategies

– Collaboration and different roles and responsibilities within the MLOPs process

Add comment

Highlight option

Turn on the "highlight" option for any widget, to get an alternative styling like this. You can change the colors for highlighted widgets in the theme options. See more examples below.

Instagram

Instagram has returned empty data. Please authorize your Instagram account in the plugin settings .

Ivana Kotorchevikj

Categories count color

Advertisement

Small ads

Flickr

  • Herr Krähenfüß
  • Sarah
  • Matty
  • on rear
  • Traffic lights
  • WoMo
  • Presentation
  • endless
  • stand still screening-smoking girl

Social Widget

Collaboratively harness market-driven processes whereas resource-leveling internal or "organic" sources.

ThemeForest