Keep in mind that it's highly unlikely that any type of extension or method of customization hasn't already been established and tested yet with a SaaS-model e-commerce hosting solution. So, it will be reinventing the wheel without the safeguards. Many of those solutions also give you the ability to build extensions and add-ons for the service as well - which can then become products themselves.
The reason I know, is that - for six years - I was the fifth employee hired at a major e-commerce solution that grew into 75+ employees and a million dollar company.
Sure, creating a inventory system and product catalog would be a nice project, but you also have to go through required steps and thousand dollar evaluations to accept some major credit cards, unless you're paying for a gateway, and that will add to the monthly expenses as well as its own set of rules.
It also needs to be secured and locked down, because if someone purchases a product or service from an engine you write from scratch, you will be personally accountable for anything that isn't done correctly - and some of those things involve serious laws about monetary transactions. and the security of the buyer's information.
With a team of five, it took a good three years for the company to grow into something with serious potential. None of it was based on a JS Framework, the servers needed to be supervised and maintained, backups alone would take over two days to be completed (as the client-base grew), the work that needed to go into the back-end started exceeding the overall intention of the original business model and front-end themes had to stick to a strict structure or it wouldn't work within the systems logic.
And these are the least of the problems they encountered. So, if this is meant to be a project that takes a few months, understand that it will never fit within that timeline.
Also, Reaction is dealing with 300+ issues with critical bugs in the list - and even though it's open-source on Github - their site conflicts with that model since it's all about marketing services and obtaining information so they can become a paid e-commerce hosting solution. That simply leads me to believe that it's not intended for anything other than a service that you can "try to get working on your own with no support", since they seem to have enough problems.
E-Commerce is a notoriously difficult world to enter, so it's best to use something that's been around for a while, is dependable, and offers developers the ability to help it grow through its APIs.
Good luck!