Whether you're trying to do automate matchmaking, or build a way to predict pricing, you may be able to build v1 in a Google Sheet, and still pass data back through to your web app! This is especially useful when it's the business and operations team that will want to refine the formulas.
At my startup, Sixty, I was in charge of building our matchmaking and quote prediction algorithms for our Hire product. I didn't know how to write it in Ruby, and it was questionable whether we could capture any of the data from users that we needed for the overall product to work.
So we built v1 in Google Sheets. The form layer was an internal tool initially. Once we proved that we could capture the data to power the algorithm and generate accurate results, we re-built the form in Vue.js on our Rails app. Then as an intermediate step, we sent the form results to Zapier as a POST request, then sent that to the Google sheet the algorithm was on.
To see the results, the Rails app triggered a Zapier workflow to look up a Google Sheets row and POST that data back into the web app. I usually don't advise repeating the overall technology setup here, but it gives you an idea of what is possible as a stopgap solution.