There isn’t a super easy way to do this from our UI (it would be easier to do from code), though it is possible. What you can do is to set up two experiment rules - the first rule targets just to the users you want to put into one of the variants, and set that split to 100%. Then add the regular experiment targeted generally. The trick here is to set the experiment tracking tracking key to the same in both rules, so the data will be tracked to the same experiment. If the number of users you want to force is high, you may get SRM warnings too.
FYI, usually not recommended for bias reasons.