This talk provides an opinionated, PostgreSQL-focused framework for application developers navigating the trade-offs between ORMs and handwritten SQL. Using examples from Django ORM, SQLAlchemy, ActiveRecord, and EF Core, it analyzes how different approaches translate into PostgreSQL queries—and how those queries interact with the planner, statistics, and indexes.
The session covers:
- How ORM-generated SQL impacts execution plans and index usage
Effects of parameterization on plan selection and performance variability
- When to use SQL in application code vs server-side functions
- Common performance anti-patterns (N+1 queries, over-fetching, excessive round-trips)

A key focus is a practical workflow for diagnosing performance issues using PostgreSQL tools such as EXPLAIN (ANALYZE), with guidance on applying precise fixes instead of trial-and-error tuning.
Designed for all application developers, this talk offers clear decision criteria and performance-oriented best practices grounded in PostgreSQL behavior.