DRY schema definitions in Postgres?

What is the recommended way to abstract out common code when writing the database schema?

I've multiple tables that have created_at and updated_at columns.

create table if not exists profiles
(
    ...
    created_at timestamptz not null default now(),
    updated_at timestamptz not null default now()
);

...

create index on profiles (created_at);
create index on profiles (updated_at);

create trigger handle_updated_at
    before update
    on profiles
    for each row
execute procedure moddatetime(updated_at);

Repeating this code for each table has the following drawbacks

1. pain to write
2. difficult to refactor (say you have to change created_at to created)
3. the real intent of the code gets buried due to too much boilerplate

Other usecases

• automatically adding an index to all foreign key constraints
• automatically creating mapping tables to represent relation b/w two entities

Solutions I'm aware of

1. use an ORM - most ORMs allow you to write abstract classes that don't map to any database table, but can be inherited by child classes that do map to tables.
2. write a function that takes in a table name and alters the table to add the desired columns, triggers and indices

   create or replace function add_created_at(table_name varchar(63))
       returns void
       language plpgsql
       security invoker
   as
   $$
   begin
    execute format('alter table %I add column created_at timestamptz not null;', table_name);
   end;
   $$;
   

   I don't particularly like this because I'm being forced to encode the query as a string.

Is there a better way?