Azure MLOps -DevOps for Machine Learning (Part 1) | Complete Intro to Azure Machine Learning Service

This Azure MLOps series contains 10 parts that show you how to apply DevOps for Machine Learning (MLOps) on Microsoft Azure.

The GitHub repo with codes I used for this series:

🔔 Subscribe for more cloud computing, data, and AI analytics videos
by clicking on subscribe button so you don't miss anything.

✅ You can contact me at:

LinkedIn:
Email:

#AzureML #AzureMachineLearning #MLOps #AzureMLOps #AzureDevOps #DevOpsForMachineLearning #MAchineLearningOperation #Azure #AI #DevOps

[vikrambit]

In this Video, MG wonderfully explains:
1.What is MLOps
2. A walk through all the components of Azure Machine Learning (previously Studio) like Notebooks, AutoML, Designer, Models etc.
3.He covered all the components of Azure Machine Learning

[myself]

🎯 Key Takeaways for quick navigation:

00:00 🌐 *Overview of MLOps and DevOps*
- Definition of MLOps and DevOps.
- DevOps as a mindset for efficient software development.
- Automation of the end-to-end software development lifecycle.
01:35 🚀 *Introduction to Azure Machine Learning Service*
- High-level overview of Azure Machine Learning Studio.
- Importance of Azure Machine Learning Studio for collaborative and governed machine learning.
- Challenges faced by data scientists in tracking and managing machine learning life cycle on personal devices.
02:33 📊 *Azure Machine Learning Studio Features*
- Quick walkthrough of Azure Machine Learning Studio.
- Collaborative workspace for data scientists.
- Importance of reproducibility and traceability in the machine learning life cycle.
06:38 ⚙️ *Using Notebooks and Scaling Compute Resources*
- Utilizing Jupyter Notebooks in Azure Machine Learning Studio.
- Creating and scaling compute resources for machine learning tasks.
- Benefits of cloud-based compute resources and scaling for efficiency.
12:14 🤖 *Automated Machine Learning (AutoML)*
- Explanation of Automated Machine Learning (AutoML) and its purpose.
- Streamlining the process of model training, hyperparameter tuning, and selection.
- Benefits of using AutoML for efficiency and model selection.
16:58 🎨 *No-Code Machine Learning with Designer*
- Introduction to the Designer tool for a no-code machine learning experience.
- Drag-and-drop approach for designing machine learning workflows.
- Automatic generation of code based on the designed workflow.
19:32 📝 *Model Registration and Deployment*
- Registering machine learning models for governance and traceability.
- Deployment of registered models as web services.
- Configuring deployment settings and authentication for model services.
20:56 💻 *IDE Integration and Cluster Computing*
- Integration with popular IDEs like Jupyter, VS Code, and R Studio.
- Terminal access to the compute environment.
- Utilizing cluster computing for parallelized machine learning tasks.
21:38 ⚙️ *Environment Configuration in Azure ML*
- Azure ML allows you to create and manage compute environments dynamically.
- The flexibility to scale the compute resources up or down based on your needs.
- Cost-effectiveness by paying only for the resources you use, providing a pay-as-you-go model.
22:04 🌐 *Environment and Deployment in Azure ML*
- Importance of environment consistency between model training and deployment in production.
- Registering environments in Azure ML to capture the operating system, packages, and general environment settings.
- Data store usage to securely store credentials for connecting to different data sources without exposing passwords.
23:01 🖼️ *Data Labeling and Linked Services in Azure ML*
- The role of data labeling, particularly in tasks like image classification or object detection.
- Utilizing Azure ML for data labeling through a user interface.
- Linking services, such as Azure Synapse, to leverage additional compute resources for distributed processing.
24:14 📚 *Conclusion and Further Learning*
- Encouragement to explore Azure Machine Learning documentation for more in-depth details.
- Emphasis on the high-level understanding gained about Azure ML capabilities in the video.
- Anticipation of the next video series focusing on Azure DevOps, building on the foundational knowledge gained in this overview.

Made with HARPA AI

[SQLTalk]

45 seconds into it and I already like it. Thank you for creating this video series.

[sabyasachighosh]

Hi, Great videos MG! A complete end to end MLOPs series.

[sirginirgin]

Kindly keep up the enthusiasm and the energy. Your way of explanation is very SIMPLE for DIFFICULT CONCEPTS. GOD bless you. Keep SMILING

[ashishsangwan]

This playlist is awesome, I would be recommending it to my friends

[shivibhatia]

Top class this is the best tutorial on MLops and intro to Azure ML. Thanks for uploading

[MrEmphonic]

One of best overviews I have seen so far!

[avicool]

Kindly post a basic video end to end, starting from creating data, create model, train model, getting data from model.

[_tyit_poojapramanik]

Content is good only the issue is that voice is too low.

[geen]

Key Takeaways for quick navigation:

00:00 🌐 Overview of MLOps and DevOps
- Definition of MLOps and DevOps.
- DevOps as a mindset for efficient software development.
- Automation of the end-to-end software development lifecycle.
01:35 🚀 Introduction to Azure Machine Learning Service
- High-level overview of Azure Machine Learning Studio.
- Importance of Azure Machine Learning Studio for collaborative and governed machine learning.
- Challenges faced by data scientists in tracking and managing machine learning life cycle on personal devices.
02:33 📊 Azure Machine Learning Studio Features
- Quick walkthrough of Azure Machine Learning Studio.
- Collaborative workspace for data scientists.
- Importance of reproducibility and traceability in the machine learning life cycle.
06:38 ⚙ Using Notebooks and Scaling Compute Resources
- Utilizing Jupyter Notebooks in Azure Machine Learning Studio.
- Creating and scaling compute resources for machine learning tasks.
- Benefits of cloud-based compute resources and scaling for efficiency.
12:14 🤖 Automated Machine Learning (AutoML)
- Explanation of Automated Machine Learning (AutoML) and its purpose.
- Streamlining the process of model training, hyperparameter tuning, and selection.
- Benefits of using AutoML for efficiency and model selection.
16:58 🎨 No-Code Machine Learning with Designer
- Introduction to the Designer tool for a no-code machine learning experience.
- Drag-and-drop approach for designing machine learning workflows.
- Automatic generation of code based on the designed workflow.
19:32 📝 Model Registration and Deployment
- Registering machine learning models for governance and traceability.
- Deployment of registered models as web services.
- Configuring deployment settings and authentication for model services.
20:56 💻 IDE Integration and Cluster Computing
- Integration with popular IDEs like Jupyter, VS Code, and R Studio.
- Terminal access to the compute environment.
- Utilizing cluster computing for parallelized machine learning tasks.
21:38 ⚙ Environment Configuration in Azure ML
- Azure ML allows you to create and manage compute environments dynamically.
- The flexibility to scale the compute resources up or down based on your needs.
- Cost-effectiveness by paying only for the resources you use, providing a pay-as-you-go model.
22:04 🌐 Environment and Deployment in Azure ML
- Importance of environment consistency between model training and deployment in production.
- Registering environments in Azure ML to capture the operating system, packages, and general environment settings.
- Data store usage to securely store credentials for connecting to different data sources without exposing passwords.
23:01 🖼 Data Labeling and Linked Services in Azure ML
- The role of data labeling, particularly in tasks like image classification or object detection.
- Utilizing Azure ML for data labeling through a user interface.
- Linking services, such as Azure Synapse, to leverage additional compute resources for distributed processing.
24:14 📚 Conclusion and Further Learning
- Encouragement to explore Azure Machine Learning documentation for more in-depth details.
- Emphasis on the high-level understanding gained about Azure ML capabilities in the video.
- Anticipation of the next video series focusing on Azure DevOps, building on the foundational knowledge gained in this overview.

[tanb]

I created my account on Azure for the very first time after watching these videos. Could you please create a video which goes one step back and explains how to create an Azure Machine Learning environment?

[elahedorani]

Really thanks to share this overview about the Azure ML Studio, but you didn't mentioned about running scripts on local computer and log them on Azure ML as well. Hope to see on your next videos :)

[vigneshnagaraj]

Please post video about how to add IDE's like vscode or pycharm to Azure ml

[PuneetMakhija-bq]

please guide which tool is used by Azure machine learning for model experimentation
right now, we are not using Azure machine learning cloud
we have our own on-premise solution where we are using mlflow for this process
but im eager to know that, which tool they are using for that
any link for this knowledge would be very helpful .

[nilayc]

Great content 👍 very useful information and Nicely articulated. Thank you.

[sivas]

Did Azure is providing certification for MLOps, may If yes may i know the certification number please .

[pankajsingla]

Hi @MG, I am creating azure mlops pipeline.
1) i have all resources but each resource have in differ RG. So can you give me tip how i can create mlops pipeline with use of these resources in differ differ RG?
2) Can i use compute instance for mlops CI/CD pipeline rather to compute cluster?

[HiteshTulsaniAtGooglePlus]

Thanks MG for these wonderful videos. Really appreciate the work. I have sent you an email with some queries - I would love to hear your thoughts on the same. Regards, HT