py-mysql-type-plugin VS powderkeg

It's fine advice - if you can type check your queries. My colleague wrote a mypy plugin for parsing SQL statements and doing type checking against a database schema file, which helps to identify typos and type errors early: https://github.com/antialize/py-mysql-type-plugin

As a data engineer, the pattern the OP shares is very familiar. I find it much preferable to use of ORMs for wide variety of reasons. However, I view implementing with SQL as an antiquated problem rather than a pragmatic feature. The evolution of this pattern would be to integrate database querying into languages more directly and eliminate SQL entirely. While this could be achieved in Python, I find that a language like Clojure, via functional programming (FP) primitives and transducers, is a natural candidate, particularly for JVM implemented databases. Rather than encapsulating SQL via query building or ORM based APIs, an FP core could be integrated into database engines to allow, via transducers, complex native forms to be executed directly across database clusters. Apache Spark is an analog of this. In particular the Clojure project, powderkeg (https://github.com/HCADatalab/powderkeg) demonstrates the potential of utilizing transducers in a database cluster context.