pregenerate schema and cache it for serverless cold start?

[vt-rc]

I am running graphene-django on serverless
I am posting here because Schema loading is in graphene.Schema.
often time there is cold start and the cold start is very slow because the schema loading is very slow, 1xx+ms for a fairly small schema. For larger service the schema load time can be over 3xx-4xx+ms

(please don't tell me to stop running serverless, thx)

Are there ways to pregenerate the schema and load from cache?
Or other way to speed up start time?

[vt-rc]

this should be open at graphql-core