Most SaaS products break at the exact moment they start working.
The customer base doubles and the cloud bill triples. The monolith that shipped fast in year one now takes an afternoon to deploy. Billing double-charges someone on a Friday night. Nobody wants to touch auth. Everyone is busy — and the roadmap stops moving.
I've spent 8+ years on the other side of that problem.
At Clinicorp (healthcare SaaS) I helped take the platform from 4,000 to 22,000 customers — 10K to 150K users — over 4.5 years, first as an engineer, then as solutions architect, then leading up to 12 developers. On the way: cloud costs down 33% while the customer base doubled, a Node.js monolith broken into microservices on GCP (Cloud Run, Pub/Sub, BigQuery, Spanner, AlloyDB), and login hardened with data fingerprinting and biometrics.
Today I build a multi-tenant clinic-management SaaS end to end: Brazilian e-invoicing (NFS-e, NF-e, NFC-e with POS), recurring payments and automatic reconciliation via Stripe and Pagar.me, and WhatsApp scheduling on the Meta Cloud API. The stack: Node.js, TypeScript, NestJS, React, Next.js, React Native, PostgreSQL, Docker and GCP, with REST and event-driven APIs. I also run AI agents as part of my daily delivery — behind a 2,500+ test suite and an adversarial review pass before anything merges.
What I'm actually good at: money-critical systems (payments, invoicing, billing) that can't lose a cent, distributed systems that stay cheap under growth, and writing the ADR before writing the code.
Two MBAs (Solutions Architecture; Cyber Security), a BSc in Software Engineering, and 3 years teaching programming at SENAI — which is why I can explain a trade-off to a non-engineer without hand-waving.
Open to Senior/Staff engineering roles, remote (LATAM · UTC-3).