Our own approach is to keep singer.io et al and JSON (flexible! streaming capable!), but deeply invest in JSON Schema to a) statically infer and map into other

Our own approach is to keep singer.io et al and JSON (flexible! streaming capable!), but deeply invest in JSON Schema to a) statically infer and map into other kinds of schema on your behalf -- TypeScript, Elastic Search, and SQL DDL so far -- and b) optimize validation so it's fast enough to be in the high-scale critical path. If you validate every document before it's read or written by your hacky transform, you've gone a long way towards limiting the bug blast radius.