Runbooks, QA lists, and architecture notes stay visible.
Savio Filho / Portfolio / 2026
Backend product engineer from Brazil
I build product systems with approval paths, boring operations, and proof you can inspect.
The work here is centered on trust layers, tenant boundaries, webhook lifecycles, upload pipelines, billing states, AI cost controls, and the docs that keep those systems understandable after launch.
Public repos ship with tests, build steps, and explicit trade-offs.
Private systems still show workflow design, constraints, and audit trails.
Why this site exists
This is a field dossier, not a gallery wall.
GitHub is where I keep the code. This site is where I explain what the system does, where it can break, and why the decisions matter once another engineer or operator has to touch it.
Operating principles
01
Approval and execution flows that make AI actions safer to ship.
02
Billing, auth, and queue-backed systems that still behave after day one.
03
Portfolio writing that shows constraints, failure modes, and operator-facing detail.
Selected work
Flagship systems with clear technical boundaries
01
VOWGRID-API
An agent-trust platform with simulation, policy evaluation, approvals, execution receipts, and rollback visibility.
The core signal is not just AI orchestration. It is the trust layer around execution: propose, simulate, evaluate, approve, execute, receipt, and rollback visibility.
02
OnboardPulse
A multi-tenant onboarding SaaS with billing, follow-up automation, storage controls, and product-like operational docs.
This project shows that I can work beyond CRUD: tenant isolation, budgeted AI usage, billing webhooks, storage policy, and cron-driven follow-ups.
03
MailSieve
A lightweight signup risk API with OpenAPI contract, key management workflow, rate limits, and deploy verification scripts.
MailSieve is a good example of a small API product where contract clarity, ops scripts, and monetization readiness matter more than feature count.
Private work, still legible
Code can stay private. System thinking should not.
The private projects here are documented with the same standard as the public ones: workflow boundaries, operational constraints, and the trade-offs that shaped the build.
Writing
Short notes written like engineering field reports.
The writing index is organized by the system edges I care about most: AI guardrails, SaaS operations, backend contracts, and a small builder-notes track.
Featured notesSimulation before executionAgent workflows become easier to trust when the system can simulate an action before it earns the right to execute it.
Browse all notesOperating map
The public stack I want recruiters to understand first
Trust + approval systems
VOWGRID-API focuses on simulation, policy evaluation, approvals, execution receipts, and rollback visibility.
SaaS workflow operations
OnboardPulse shows tenant isolation, billing, follow-up jobs, upload flows, and AI budget edges.
Lean product APIs
MailSieve and the supporting notes show how narrow products still need contracts, verification, and ops discipline.