Design for a Cloud Native Application in Azure for ML Insights and Actions

Question

I have a plan to create a cloud-native algorithmic trading application, ideally using only PaaS and SaaS (no IaaS), and I'd need some input on how I should go about doing it. The idea is quite simple in that I want to use an API query to get financial trading data from an external SaaS solution, feed it into several Azure PaaS products (most notably ML for modelling), and then do something with it. I've come up with the following high-level diagram so far:

Review of the Solution

Although I am familiar with Azure, I am not an Azure cloud engineer and have little hands-on experience developing solutions. I then want to use this endeavour as a foundation to educate myself more.

I debated whether or not to use Event Hubs as soon as I started working on the build. Conceptually, that makes sense because I'm separating how a data stream is produced from how it is used. This should make it simpler for me in the future if I need to alter the data feed(s). I also considered the location of the data storage—should it be an Azure Table or a SQL database, to put it simply? The rationale behind this is that as I loop through my models, the trading data will need to be saved for regression testing. With that said, I'm hoping someone with experience in this field may share some advice.

Thanks!