What is most valuable?
I wouldn't say it's necessarily about liking everything about the platform entirely. It's more about what do we want? In terms of machine learning, there are times that we have to get into it and customize it, etc. We can use the ready-made models that are available without really having to code encrypt them with our bitcoin code — our model doesn't need to be too complex. Deployments and everything, in general, can be automated from a CI/CD perspective as well.
What needs improvement?
I really can't see where it needs much improvement. My experience is only half-matured and is still maturing.
I don't think we have reached the stage where the customer has enough cohesion to really complain about anything. Also, a Microsoft team is personally involved which really simplifies the process.
In the machine learning world, when you are defining the model, typically people go for an interesting library of algorithms that are available. It's an imperfect scenario. The world is not as ideal as we think: how we draw a mathematical or theoretical formula is not exactly as it seems. With encryption, this uncertainty is actually much higher — that's why you need to tweak your mathematical formula or completely customize it. For this reason, my team has a development platform where they can customize code when it fails.
For how long have I used the solution?
I have been using this solution since June.
What do I think about the stability of the solution?
Regarding the stability and scalability — so far so good; however, we're still exploring quite a bit. It's too early to really comment because the customer has already paid. They've just started their journey. We are yet to explore exactly what and how they want to use it.
How are customer service and technical support?
So far, we haven't had a situation where we have needed to raise a ticket for support on a technical front.
Currently, we're handling any issues internally because we're still in the initiation stage. It's going to take some time for us to really get our hands into it, but so far it's been a really good experience. Based on various conversations that I was part of, I think our customer really appreciates the support coming from our people.
How was the initial setup?
Compared to similar solutions, Microsoft Azure Machine Learning Studio is quite new so the initial setup wasn't much of a challenge. The data processor can pose a bit of a challenge, but the real complexity is determined by the skill of the implementation team.
What other advice do I have?
I would Definitely recommend Azure Machine Learning Studio — no doubt about it, it's a full-contact solution. Having said that, it really depends on the customer's appetite and what they're comfortable with. For example, I have interacted with people who prefer a basic Google cloud platform — from an AML perspective, they just feel like it's primarily Google. Not because of AML per se, it's more from a data storage perspective, which in this case, works better.
Personally, I come from a VFA site in the financial sector. Over there, the customers are really conscious about hosting their station or their data, especially on the cloud. Typically, they are very restricted because they are not comfortable hosting customer data on the cloud. This is where I think Azure or Google or even AWS fall short — they don't play any role there. Because of this, people actually customize their solutions or model them to fit their custom sites and customer-based solutions.
Overall, I would give this solution a rating of seven. It's a great option if you are fairly new and don't want to write too much code. As long as the model is not too complex, it's a pretty easy solution to roll out.