Systems engineer. PostgreSQL, AWS, and the parts of production that fail quietly.
Most monitoring tells you when something is down. I build tools for the other case: the plan that drifted, the index that stopped returning what it used to, the constraint your schema claims and never enforced. No error, no alert, no log line — just a system that is no longer doing what you believe it does.
Five, all published on PGXN. All pure SQL, no shared library, so they install on RDS, Aurora and Supabase, where extensions written in C cannot go.
| Extension | What it watches |
|---|---|
| pg_plan_guard | Detects when a query plan drifts away from the plan you approved. Built on pg_plan_advice, new in PostgreSQL 19. |
| pg_recall_guard | Watches vector indexes for recall drift against a baseline you approved. Works with pgvector, pgvectorscale and any index with distance ordering operators. |
| pg_promise_guard | Finds the guarantees your schema claims and silently does not enforce: invalid unique indexes, NOT VALID constraints, disabled triggers, RLS without FORCE. |
| pg_grammar_guard | Compiles a token-level grammar from your live catalog, and watches it for drift. |
| pg_living_assertions | A registry of things you claim are true, with the SQL that proves them and the date they were last proven. |
Every one of them ships a regression suite that runs with make installcheck, and each
was tested against more than one PostgreSQL major version before release.
- Open pull request on pgvector.
- BUG #19638 reported against the planner.
- Ported pgvectorscale to the PostgreSQL 19 beta. It runs in production here, behind 30 vector and full-text indexes in daily use.
Around twenty applications on AWS across three accounts, on PostgreSQL 17 and 18: multi-tenant Row Level Security, logical replication, backup and restore, query and schema design, and SQL Server to PostgreSQL migrations.
n8n community nodes published on npm — LATAM identity, shipping and e-signature, WhatsApp, Square, Stability AI — and a fork of llama.cpp with an NPU backend.
Remote from Chile, with daily overlap with US hours. Performance and query work, schema design, migrations, extension development, and fixed-price audits of slow queries or AWS cost.