I'm working at getting Scientist utilized within the GitLab project because I think we can use it successfully in a couple of ways. (👋 hello fellow Git* person). One way is the standard usage, and the other is in a slightly more A/B/n capacity, which turns out is only subtly different in terms of how Scientist is implemented at a high level.

The main difference is in how one may want to generate results of an experiment. Currently Scientist has the concept of running all variants and comparing them in terms of performance and verifying that results match. This is obviously highly useful and will always be the primary usage, but a different approach could be to run a subset of variants, or a single variant -- let's say using Unleash behind the scenes to determine which one(s), and the rollout strategy.

I didn't change or add any tests because the behaviors are functionally unchanged, but I'd be happy to if there's any guidance in how you'd like to see it be tested.

Thanks for the work on maintaining this project!