What is our primary use case?
We used this solution for defining new predictive models, such as recommendation systems, but also price elasticity models for fraud detection, and the classification of customers.
We are not using this solution regularly. We are now using Azure Databricks.
What is most valuable?
The most valuable feature of this solution is the ability to use all of the cognitive services, prebuilt from Azure. You just have to drag and drop the services into your pipeline, and it can be applied through the pipeline. It's very helpful for data scientists. If you don't have any special knowledge in data science, just to know that you want to consume a service, that's all you need.
They have a tool for data gathering from some social networking sites such as Twitter and Facebook, which is great.
What needs improvement?
If you want to be able to deploy your tools outside of Microsoft Azure, this is not the best choice.
One of the problems that we had was that you could only execute the model inside the machine learning environment. Comparing this to Databricks, if you create a pipeline, it could be in a notebook and you have all the code and then you can export your notebook to some other tool directly, for example in Jupyter and Spark. If you change tools then you won't lose your assets.
I would like to see improvements to make this solution more user-friendly.
They need to have some tools, like Apache Airflow, for helping to build workflows.
Better tools are needed to bring the data from existing storage into the environment where they can play with it and start to analyze what they already have, on-site. This is what the majority of people would like to do.
A feature that would be useful is to have some standard data transportation functions. They have ADF, Azure Data Factory, but it's a little bit heavy to manipulate. If they could have something more user-friendly, like Apache Airflow, it would be very nice.
For how long have I used the solution?
We have been using this solution for almost nine months.
What do I think about the stability of the solution?
In general, there are no major issues.
What do I think about the scalability of the solution?
We never went into production because we switched to Azure Databricks. We did, however, try some performance testing and tried scaling some resources. The scalability of this solution is quite easy.
It is not difficult compared to some of the other tools that are available on Azure.
We have only five users including data engineers, data scientists, and one data DevOps engineer who was working with us on creating all of the DevOps pipelines for deploying all of our models.
How are customer service and technical support?
I have been in touch with technical support many times. The client I work for is a first-year client for them and we received some very useful support. The showed great willingness to help and they provided a lot of support for free.
We also had meetings with some experts on their data side and we had some free consultancy days given by Microsoft. It is called FastTrack and it is only available for some kinds of clients.
We are completely satisfied with the technical support.
Which solution did I use previously and why did I switch?
We did not use another solution prior to this one, but we now use Azure Databricks.
How was the initial setup?
The initial setup of this solution is straightforward.
Besides that, once we fixed our network problem, it was straightforward.
What about the implementation team?
We implemented this solution on our own. The documents available on Microsoft Online made it quite easy.
What's my experience with pricing, setup cost, and licensing?
When we started using this solution, our licensing fees were approximately €1,000 (approximately $1,100 USD) monthly, but it was fluctuating. When we got our first models and were ready for the user acceptance testing, our licensing fees were between €2,500 ($2,750 USD) and €3,000 ($3,300 USD) monthly. It was quite limited.
We expected the rate to be higher than this, at perhaps €10,000 (approximately $11,000 USD) per month, but it wasn't the case.
What other advice do I have?
Microsoft has increased the usability and the features since we first implemented this solution.
If I had to start this process over again, I would involve Microsoft earlier because they were great for providing support, as well as guidance on the architecture and what kind of stuff you can do with the tool, and what you should do with it. This was very helpful to orient the team to the right documentation and tutorials.
The second thing I would do is to start working with DevOps activity as soon as you can. We found ourselves redoing the same things many times, instead of having a DevOps pipeline to implement the stuff that we already stabilized, for example, and then not losing time.
The third thing is involving an integrator to help put together the big picture.
I would rate this solution a seven out of ten.
Which deployment model are you using for this solution?
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?