[go: up one dir, main page]
Scribd
Upload
274 views1 page

Best Practices For Machine Learning Operations (MLOps)

The document discusses best practices for MLOps using Azure Machine Learning. It recommends unifying data science and engineering teams to reduce model development time from months to weeks. With MLOps on Azure, you can quickly realize value from models and maintain that value over time by creating reproducible models and training pipelines, simplifying deployment and monitoring, and automating retraining processes. It provides a 5-step process for optimizing the MLOps lifecycle including observing data drift, creating/retraining models, registering models, creating scoring files, and deploying and monitoring performance.

Uploaded by

Kabachu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
274 views1 page

Best Practices For Machine Learning Operations (MLOps)

The document discusses best practices for MLOps using Azure Machine Learning. It recommends unifying data science and engineering teams to reduce model development time from months to weeks. With MLOps on Azure, you can quickly realize value from models and maintain that value over time by creating reproducible models and training pipelines, simplifying deployment and monitoring, and automating retraining processes. It provides a 5-step process for optimizing the MLOps lifecycle including observing data drift, creating/retraining models, registering models, creating scoring files, and deploying and monitoring performance.

Uploaded by

Kabachu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Best practices for MLOps with

Azure Machine Learning

Models differ from


code because they have an By unifying our tech stack and bringing our
organic shelf life and will engineers in Big Data and online software together
deteriorate unless maintained. with data scientists, we got our development time
Once deployed, they can add
real business value, and this gets down from months to just a few weeks.
easier when data scientists are Naeem Khedarun
given the tools to adopt Principal Software Engineer, ASOS

standard engineering practices.

With MLOps on Azure


you can quickly realize the value
of your models and retain that Create reproducible Simplify model packaging, Explain & observe
models and reusable validation, and deployment model behavior and
value over time. training pipelines. for quality control, A/B automate the
testing, and more. retraining process.

5 Best Practices to optimize Create models with reusable ML pipelines using the Azure
your MLOps lifecycle on Azure: Machine Learning extension for Azure DevOps. Store your code in
GitHub so it automatically integrates into your MLOps pipeline.

1
Observe data
drift and feed 5
back model
information to
improve future
Automate your
training. Create / Retrain Model MLOps rollout
2 using Azure
DevOps + Azure
Monitor Register Model Machine Learning
Deploy and
monitor 4 MLOps for version models
with rich metadata
performance
so you can
lifecycle and event
management.
release models Web Analytics IoT
Service Edge Create Scoring Files
with confidence and Dependencies
Use Model
and know when
to retrain.

3
Automatically create an audit trail for all
artifacts in your MLOPs pipeline ensure asset
integrity and meet regulatory requirements.

Get started today Get the free trial of Azure


Machine Learning and pair it
with Azure Machine Learning with the Azure AML extension
within DevOps
Learn more

You might also like