That link doesn't support your case. PostgreSQL beat SQLite on every metric he tried. And he kept PostgreSQL's hands tied behind it's back by using only the intersection of features via an ORM and barely touching performance tweaks.

SQLite is a weird mutant. It is an RDBMS SQL database designed to be a persistence layer of a single application. It doesn't work as the former (not a server) nor is it well suited for the latter (kv databases work better as I have recently learned through experience). I think they did a great job of building the mutant to be the best mutant possible, but I don't see the use case anymore.