Microsoft Azure Machine Learning Studio Reviews

I've had experience working for two distinct companies. My previous employer operated in the telecom domain, primarily focused on telecom-related projects. In my current role, which is in the shipping domain, we primarily manage shipping cases. Furthermore, our current work predominantly revolves around a machine learning platform implemented on storage systems.

In the Machine Learning Studio, particularly the Designer part, which is essentially Azure's demo designer, there is room for improvement. Many customers and users tend to switch to Microsoft Azure Multi-Joiners, which is a more basic version, but they do so internally.

One area that could use enhancement is the process of connecting components. Currently, every time you want to connect a component, such as linking it to your storage or an instance like EC2, you have to input your username and password repeatedly. This can be quite cumbersome. Google, for instance, has made it more user-friendly by allowing easy access for connecting services within a workspace. In a workspace, you can set up various resources like storage, a database cluster, machine learning studio, and more. When connecting these services, there's no need to enter your username and password each time, making it a more efficient process. Another aspect to consider is the role of the designer, and they were to integrate a large language model to handle various tasks, it could significantly enhance the overall scalability and usability of the platform.

I have been working with Microsoft Azure Machine Learning Studio for three years. 

I would rate it eight out of ten. 

I would rate it eight out of ten.

Yes, I have worked with Azure in my previous experiences.

There isn’t any such expensive costs and only a standard license is required. 

I use Microsoft Azure Machine Learning Studio as a component in more complex stories involving Azure AI and Azure integration with business needs.

I can smoothly integrate with other Azure services, which is a valuable feature. The platform as a service provides user-friendly instruments, making the experience easy. 

Additionally, the integration with Azure services enhances workflow and meets my expectations.

The pricing policy should be improved. I find the pricing to be not a good story in this case, as it is not affordable for everyone.

I have been using Microsoft Azure Machine Learning Studio for about two years.

Stability is high, with a rating of nine out of ten.

Microsoft Azure Machine Learning Studio is scalable.

Microsoft technical support is rated a seven out of ten. There used to be better times for Microsoft technical support.

Neutral

The initial setup involves provisioning and configuration rather than installation. It is a short story, taking about ten minutes per service for proper configuration.

My team of specialists handles the implementation. We are a system integrator with a team of around two hundred engineers, twenty-five of whom work in AI and data.

Measuring ROI is hard due to the nature of projects, however, Microsoft provides a business case builder to help estimate potential profits and values from its products.

Pricing is considered to be top-segment and should be improved. I rate the pricing as three or four on a scale of one to ten in terms of affordability.

I would recommend trying and using it as it will be okay. I rate the overall solution eight out of ten.

We used this product as implementers. For example, a client wanted to use the Azure stack, specifically Azure Machine Learning Studio. Microsoft was a consultant on the project, and we were the implementation partner.

There are actually two tools. One is Azure Machine Learning Designer (which used to be called Azure Machine Learning Studio), and the other is Azure Machine Learning. Designer is a drag-and-drop interface, primarily for those without extensive coding expertise.

Azure Machine Learning has become the de facto product, and it allows you to write code and provides numerous components for building machine learning models.

We build all our machine-learning models using ML Studio. That's the primary purpose for us. We build standard models that use data and various algorithms.

My expertise was focused on the analytics aspect. We worked in a large firm with different teams responsible for MLOps and integrating outputs with other systems. 

My responsibility was to get data, build models, and provide the output. How the output was rendered and the process of doing so was handled by other teams. 

Additionally, there's a separate team that handles the productionalization of machine learning models. It's a diverse team structure, so the scope of the totality of the integration process varied. Essentially, they take the model I build, use Azure DevOps for versioning, and then deploy it to production.  

Everything in Azure is very intuitive. As a Microsoft product, it's designed that way. You can easily search for things, and many features are built-in. Many things can be easily done via drag-and-drop.

We often use custom code that we can template or create as boilerplate code for different teams. It sounds difficult, but Azure makes it very easy. 

The learning curve is very low. Operationalizing the model is also very easy within the Azure ecosystem.

The drag-and-drop interface specifically improves our workflow a lot. Designer simplifies the process when we want to create a quick model. It makes things very, very easy. It's a great starting point.

When we want to create models, model management and deployment are crucial. We had boilerplate code ready, and Azure made it significantly easier to deploy using its services. That was the easy part.

In terms of data capabilities, if we compare it to Google Cloud's BigQuery, we find a difference. When fetching data from web traffic, Google can do a lot of processing with small queries or functions. Azure didn't have that same inbuilt feature for website traffic or analytics, unlike Google DB and BigQuery.

I started using it around the year 2020. We used it up to 2023, so roughly two years.

It is a stable product. I would rate the stability a nine out of ten. 

It is a scalable solution because it works well for large databases.

I would rate the scalability a nine out of ten. 

The customer service and support are good. 

Positive

We've used Google Vertex AI and AWS SageMaker.

I was not involved in the deployment process. But there is maintenance. It was quite a headache. 

Maintenance does require attention. With any cloud implementation, cost optimization is a major factor. Our team had discussions about it. 

I would rate the pricing an eight out of ten, with ten being very expensive. Not very expensive, not very cheap.

It was on a yearly basis, and there were also usage-based costs.

If it's my recommendation, it's a very competent product. It has all the necessary features for data engineering, data science, and model management. 

It's a complete suite of products that can address your end-to-end data needs.

Overall, I would rate the solution a nine out of ten. 

The use cases of this product are primarily for the BFSI; digitization and building machine learning models that provide recommendations for creating analytical insights from extracted data. We also do Jupyter Notebook authoring. We are partners with Microsoft and I'm a practice director.  

The product enables quick data preparation and data processing pipeline as well as modeling work and it's all part of Azure Machine Learning. It also gives us an idea of what machine learning model is good to use because the hyperparameter tuning is done automatically which saves us time and effort. 

Auto email and the studio are great features. 

It's not that easy to master the program, it requires some specific learning. If we want to extend the program to include inexperienced users, it can take some time for them to learn the solution. It would be nice if they added GPU solutions. Most of the solutions coming out now are video analytics or edge computing-based and Azure should have that focus.  

We haven't had any issues with stability. 

We haven't faced any challenges with scalability. If there are any issues, our Microsoft infract team pitches in but we haven't had any serious problems. We have around 25 to 30 customers accessing this solution. Maintenance is straightforward and doesn't require more than one person. 

Customer support is very good, they are prompt and helpful in solving problems. 

Positive

Our switch to AMLS was an organic development that came from the needs of our customers and was based on the quick time to develop and the pre-built machine learning models that the solution has.

The initial setup is straightforward with deployment time depending on the environment. It depends on how many machine learning models we need to develop, the type of resources, the different sources, data volumes, etc. 

We don't deal with licensing, that is something our customers are responsible for.  My understanding is that the cost is $50 for the digitization of 1,000 pages. I think it should be reduced to somewhere between $20 to $30 per 1,000 pages so that we can make a better offer to our customers. 

I believe Azure Machine Learning has a very good pre-built model which enables quick development of solutions, particularly text analytics and cognitive-based solutions. 

I rate this solution nine out of 10. 

In terms of what I found most valuable in Microsoft Azure Machine Learning Studio, I especially love the designer because you can just drag and drop items there and apply the logic that's already available with the designer. I love that I can use the libraries in Microsoft Azure Machine Learning Studio, so I don't have to search for the algorithms and all the relevant libraries because I can see them directly on the designer just by dragging and dropping. Though there's a bit of work during data cleansing, that's normal and can't be avoided. At least it's easy to find the relevant algorithm, apply that algorithm to the data, then get the desired output through Microsoft Azure Machine Learning Studio.

I also like the API feature of the solution which is readily available for me to expose the output to any consuming application, so that takes out a lot of headache. Otherwise, I have to have a developer who knows the API, and I have to have an API app, so all that is completely taken care of by the Microsoft Azure Machine Learning Studio designer. With the solution, I can concentrate on how to improve the data quality to get quality recommendations, so  this lets me concentrate on my job rather than focusing on the regular development of APIs or the pipelines, in particular, the data pipelines pulling the data from other sources. All the data is taken care of and you can also concentrate on other required auxiliary activities rather than just concentrating on machine learning.

As for the areas for improvement in Microsoft Azure Machine Learning Studio, I've provided feedback to Microsoft. My company is a Gold Partner of Microsoft, so I provided my feedback in another forum. Right now, it is the number of algorithms available in the designer that has to be improved, though I'm sure Microsoft does it regularly.

When you take a use case approach, Microsoft has done that in a lot of places, but not on the Microsoft Azure Machine Learning Studio designer. When I say use case basis, I meant recommending a product or recommending similar products, so if Microsoft can list out use cases and give me a template, it will save me a lot of time and a lot of work because I don't have to scratch my head on which algorithm is better, and I can go with what's recommended by Microsoft.

I'm sure that isn't a big task for the Microsoft team who must have seen thousands of use cases already, so out of that experience if the team can come up with a standard template, I'm sure it'll help a lot of organizations cut down on the development time, as well as going with the best industry-standard algorithms rather than experimenting with mine.

What I'd like to see in the next version of Microsoft Azure Machine Learning Studio, apart from the use case template, is the improvement of the availability of libraries. Microsoft should also upgrade the Python versions because the old version of Python is still supported and it takes time for Microsoft to upgrade the support for Python. The pace of upgrading Python versions of Microsoft Azure Machine Learning Studio and making those libraries available should be sped up or increased.

I've been working with Microsoft Azure Machine Learning Studio for nearly two years now.

Microsoft Azure Machine Learning Studio is a stable solution. My company is already using it in production. At least customers use the recommendations from Microsoft Azure Machine Learning Studio in production, so the solution is quite stable, at least in cases developed by my company.

Microsoft Azure Machine Learning Studio is a solution that's easy to scale. It's pretty easy because it is hosted on Kubernetes, and there is an option in the portal where I can simply move my plan from standard to enterprise. The solution also has an automatic scaling option available because it is on Kubernetes, so it can scale automatically. I'm seeing that it's quite scalable. This has nothing to do with availability because it just runs in the background, and it is not customer-facing, but the output is customer-facing, so availability is a different case, but in terms of scalability, Microsoft Azure Machine Learning Studio is scalable.

The technical support team for Microsoft Azure Machine Learning Studio was pretty good, though I had to tailor the answers to my requirement, but would rate support a four out of five. Most of the questions my company had, more or less, the support team already experienced, so the team had answers readily available which means there wasn't a need to do a lot of R&D, so getting answers from technical support didn't take a lot of time.

In terms of setting up Microsoft Azure Machine Learning Studio, initially, when my company started, the documentation wasn't so good, but now it has improved. Provisioning the solution only takes a few clicks, so it's no big deal, but setting up the pipelines because no enterprise will have a single environment, you'll have to create multiple pre-production and end production environments, so moving my latest changes to the next environment was a bit of a challenge.

Many terminologies are now in the market such as DevSecOps, and MLOps, so that MLOps documentation was available initially, but it wasn't very explanatory, but now, there's a lot of improvement in the MLOps documentation and that will help me move and propagate my changes from one environment to another.

Microsoft has made improvements into the tutorials, especially on MLOps. Finding MLOps experts in the market was also very tough initially, so my company was trying to learn on the job and do it, so it took some thinking and time, but it's still good because you can learn on the job and do it, but you won't always have the luxury of time to learn it.

In terms of pricing, for any cloud solution, you should know the tricks of the trade and how to use it, otherwise, you'll end up paying a lot of money irrespective of the cloud provider, so at least for Microsoft Azure Machine Learning Studio pricing versus AWS, I would rate it three out of five, with one being the most expensive, and five being the cheapest. It could be cheaper, but you also have to be careful when choosing the plans, for example, consider the architecture and a lot of other factors before choosing your plan, if you don't want to end up paying more. If your cloud provider has an optimizer that seems to be available in every provider, that would keep alerting you in terms of resources not being used as much, then that would help you with budgeting.

We evaluated quite a lot of options. We compared Microsoft Azure Machine Learning Studio against Google Cloud and AWS solutions, and there were several others available in the market. I'm trying to recollect the names which we compared the solution with. We did the benchmarking, but we went with Microsoft Azure Machine Learning Studio because our clients and their data were on Azure, though that doesn't necessarily make you go with the solution. After all, you can pull the data from any other cloud as well. For our use case, however, we found many of the things were readily available and the learning curve for Microsoft Azure Machine Learning Studio compared to others was better and easier. We didn't have to search for experts in the market to hire them because we could have our in-house team learn and deliver the solution on the job.

Microsoft Azure Machine Learning Studio is a cloud-native solution. It's completely cloud-based.

My company has eight users of Microsoft Azure Machine Learning Studio.

My rating for Microsoft Azure Machine Learning Studio is seven out of ten.

We primarily use the solution for our projects. We're currently adopting it for a competitive, new initiative. We create a lot of products related to AI inside the organization as needed for business cases and different business deals. We use it for data extraction and language processing. 

They've been helpful with hands-on experience.

It's easy to use.

The deployment is fast.

The interface has been very good so far. 

It has good configurations. 

It's stable.

The solution scales well.

There have been issues with environmental creation. It can take a lot of time. The speed of deployment should be faster, as should testing. 

We've been using the solution for about six months. It is fairly new. 

It is a stable solution. It's reliable. I'd rate its stability ten out of ten. There are no bugs or glitches, and it doesn't crash or freeze. 

The solution is scalable. I'd rate it ten out of ten. 

We have 10 to 15 users as of now on the product.

We use it often. 

We can deploy the solution within ten minutes. 

There is a team that handles the deployment. 

We don't have to really worry about maintenance; we're still in the process of adoption.

Our team handles the deployment in-house. 

We are customers and end users.

We're using the latest version f the solution. 

I'd rate the solution ten out of ten. 

We primarily use the solution for sales forcasting and for creating a pipeline in Azure. We are publishing the pipeline from Azure DevOps, and through the AML endpoint so that the pipeline will run one after the other models. These predictions will be stored and we can visualize everything. 

The designer and notebooks are great. We like the pipelines we are able to deploy and the process is very simple.

The visualizations are great. It makes it very easy to understand which model is working and why.

The setup is simple. 

It is stable and reliable.

I have had no trouble scaling.

Technical support is good. 

The solution cannot connect to private block storage. It does not allow this connection, which is a pain point. The confidential data needs to be removed from the block, and that becomes a security issue. 

In Azure Databricks, how we are promoting the models could be easier. The UI in Daabricks is a bit easier. We'd like ML Studio to be streamlined. 

I've used the solution for about two and a half years. 

The solution is stable and reliable. There are no bugs or glitches. It doesn't crash or freeze. The performance is good. 

The solution can scale. I haven't used Azure Kubernetes services yet. However, I haven't had issues with scaling so far. 

We have around ten to 20 people on our project using the solution. Many users use it in our company - not just on my team.

I've reached out to technical support. They have SLAs in place that help us to troubleshoot issues. Even critical issues get sorted out quickly. We're using premium Microsoft technical support. 

We also use Databricks. In Databricks, there is no designer module to design pipelines. There are other features available. 

They do behave in the same way; however, in Databricks, I do need to do more configurations and a bit more work with it. Still, it allows me to connect to private blocks, which I cannot do in this product. It also requires me to run job clusters separately. 

Security-wise, Databricks is more secure. 

This is easy to deploy. I did not fid the process to be overly complex. 

The solution has a higher price. I'd rate it three out of ten in terms of affordability. 

I am an end user. 

I'd rate the solution eight out of ten. I'm pretty happy with its capabilities. 

Developing and operationally implementing a powerful lead scoring model for a major Multiufamily developer and operator of apartment properties throughout major western states. The work included 3 years of data across over 60 properties with more than 500,000 leads and 3 million transactions.

Increased sales force productivity by permitting them to prioritize activity during peak leasing periods on those leads most likely to close

The ability to do the templating and be able to transfer it so that I can easily do multiple types of models and data mining is a valuable aspect of this solution. You only have to set up the flows, the templates, and the data once and then you can make modifications and test different segmentations throughout.

We were working across a number of internal departments as well as some outside departments and this solution made it extremely easy to communicate across functional area because it was all in flow chart and data form so that if somebody had an issue, like changing the data set or something like that, they could point right to it and we could get that handled and incorporated into the model. It's extremely efficient on the computer. We had to do a number of resets on the data in the model and to be able to turn things around and validate the model and the new set in two hours, was just incredible for me.

It was very robust. The ability to move the objects around so easily and then communicate is really its power. Then to be able to show it to the sales and senior management, in terms of what was employed and made it very easy to get my job done.

In terms of improvement, I'd like to have more ability to understand the detailed impact of the variables on the model and their interactions. Their algorithms are very powerful and they explain overall the net contribution of each of the variables to the solution. In terms of being able to say to people "If you did this, you'll get this much more improvement" Azure (at least my understanding of it) doesn't provide readily accessible tools to assess from a management perspective the impact of their changing a sinimized, the better.gle value - for instance in closing a lead, decreasing response time by 10%.

I recognize that the multivariate algorithms used from decision trees to neural nets do not readily provide the coefficients for each variable ala the older regression modeling approaches. My experience over my 50 years of developing and implementing predictive models has been that more than half the value of modeling lies in improving management's understanding of the process being modeled, often leading to major organization and operational structure changes. More ability to understand the variables impacting the end result being optimized would be very useful. 

I have worked extensively with this solution for the last three years. 

I haven't had any problems with stability. 

I didn't have any issues with the scale. we rapidly went from test to full implementation across all datasets.

I never had to use technical support.

I have used SPSS modeler (part of WATSON really) but because client was a Microsoft shop, I switched to Azure.

I found the setup to be very easy. I've been doing this type of work for 50 years so the modern terminology isn't always the same as what I grew up with. It took me a while to understand that, but the setups were very easy. As with anything, the hardest part is always getting the data together, but the outside consultants had built up a very, very good data warehouse. The ability to manipulate the data and create variables was very nice.

THIS IS THE ONLY MODELING APPROACH THAT EVER WORKED THE VERY TIME I RAN IT!!

Because client isa Microsoft shop, everything was Microsoft in terms of having solutions like Power BI and stuff like that. Azure is very useful and very inexpensive.

The major advice I give is that clients must get the user,somebody who understands the business issues, to be deeply involved with it and the data transformation. Most people don't. And that's true for data science applications. We don't just follow the data in a big pile and remodel, we advance the process that we're modeling. Consider what transformations of the data you need to make it workable and usable.

Remember, over half the initial value of modeling is the strategic understanding provided re the importance of different variables to the model and hence the organizaion's performance. Very often the modeling identifies opportunities for changing structures, decision rules, etc. even prior to the model's actual implementation technically.

I would rate it a nine out of ten.