Software & Applied AI Engineer
Backend Engineering · Applied AI · Cloud & Serverless · Product Engineering
Software engineer focused on backend systems, enterprise applications, product engineering and Applied AI.
My main background is in TypeScript, Node.js, NestJS, PHP, Laravel and relational databases, working with APIs, enterprise integrations, business-critical systems, legacy modernization and production environments.
I also design and deploy applications using cloud and serverless platforms, including Vercel, Supabase, Neon, MongoDB Atlas and AWS.
AI is part of my daily software engineering workflow — across personal, enterprise and public-sector projects.
| Area | Technologies |
|---|---|
| Backend | TypeScript · Node.js · NestJS · PHP · Laravel · .NET |
| Frontend | Next.js · React · Angular |
| Mobile | Kotlin · Jetpack Compose · React Native |
| Databases | PostgreSQL · Oracle · SQL Server · MongoDB · MySQL |
| Messaging & Cache | RabbitMQ · Redis |
| Cloud & Serverless | Vercel · Supabase · Neon · MongoDB Atlas · AWS |
| Infrastructure | Docker · Nginx · PM2 · CI/CD |
| Architecture | REST APIs · Microservices · Monoliths · Serverless · PWA · Legacy Modernization |
AI is not just a code completion tool in my workflow.
I use coding agents and LLM-powered tools to investigate systems, understand business rules, plan changes, implement features, review code, validate solutions, document decisions and evolve products.
Claude Code · OpenAI Codex · GitHub Copilot · Antigravity · Ollama · Local LLMs
The tools change.
The engineering methodology around them is the important part.
One of the core practices I adopted is creating a Project Constitution.
It acts as the highest-level reference for both humans and AI agents, defining the principles and constraints that should remain stable while the product evolves.
Project Constitution
│
▼
Architecture & Business Rules
│
▼
Project Context
│
▼
Specialized Skills
│
▼
AI Agents
│
▼
Investigate → Plan → Implement → Validate
│
▼
Human Decision
A Constitution can define:
Architecture Principles · Product Principles · Business Invariants · Security Boundaries · Engineering Standards · Decision Criteria · Evidence Requirements
The goal is simple:
Prevent an agent from solving an isolated task correctly while making the overall product worse.
| Practice | Application |
|---|---|
| Project Constitutions | Persistent product and engineering principles |
| Custom Skills | Reusable knowledge and workflows for specialized tasks |
| Context Engineering | Structuring what an agent needs to understand before acting |
| Repository-aware Agents | Agents working directly with real repositories and documentation |
| Agentic Workflows | Investigation → planning → implementation → validation |
| Business Analysis | Understanding and validating complex domain rules |
| Architecture Analysis | Evaluating changes against the broader system |
| Code Review | Finding inconsistencies, regressions and implementation risks |
| Documentation | Keeping technical and product knowledge close to the code |
| Model Evaluation | Comparing models for reasoning, coding, context and tool use |
| Local AI | Running and evaluating local models and inference configurations |
| Human-in-the-loop | Keeping important architectural and product decisions human-owned |
| Project | Focus | Stack / Topics |
|---|---|---|
| HR Workflow | Recruitment and employee workflows with enterprise integrations | NestJS · Next.js · SQL Server |
| GIP Promo | Commercial and promotional processes integrated with SAP | Backend · SAP · APIs |
| TMS | Modernization of legacy transportation services | .NET · Node.js |
| SAP REST / RFC | Modern API layer for SAP integrations | NestJS · REST · RFC · SAP |
AI-assisted engineering workflows were also incorporated into the development process, including structured context, project-specific rules and agent-assisted analysis and implementation.
| Project | Focus | Stack / Topics |
|---|---|---|
| Tariffs Management System | Maintenance, features, releases, CI/CD and external integrations | PHP · PostgreSQL · SEI |
| SGSE — Ethical Sanctions | Modernization of a legacy application and government integrations | Laravel · Oracle · Gov.br · APIs |
| InfoAP — Private Archives | Migration and modernization of a legacy PHP application | Laravel · Oracle |
I apply the same AI-assisted engineering approach across these environments when appropriate, adapting constitutions, skills, context and agents to each codebase and domain.
FlowHub became one of the most important projects in the evolution of how I work with AI.
It became a practical environment for experimenting with:
Project Constitutions · Custom Skills · Specialized Agents · Context Engineering · Product Analysis · Architecture Analysis · Agentic Workflows · Validation
The main question gradually changed from:
Can AI write this code?
to:
How can an AI agent reliably contribute to a real software product while understanding its architecture, business rules, context and constraints?
Many of the practices I now apply across other projects evolved from what I learned building and evolving FlowHub.
I've also worked on business automation and conversational systems, including WhatsApp-based workflows and integrations.
NestJS · React Native · APIs · WhatsApp · Automation · Integrations
I work with both traditional infrastructure and modern managed platforms, choosing the approach according to the product instead of forcing a specific architecture.
Vercel · Supabase · Neon · MongoDB Atlas · AWS
Serverless Functions · Managed PostgreSQL · Managed MongoDB · Authentication · Storage · Realtime · Edge · CI/CD
I particularly enjoy architectures where managed services allow a small team to build and operate products without unnecessary infrastructure overhead.
What started as motorcycle maintenance eventually became a long-term engineering and experimentation project.
I document modifications, reasoning, tests and observed results while exploring areas such as:
Intake · Airflow · Carburetion · Ignition · Rotating Mass · Exhaust · Transmission Ratios · Wheels & Tires · Street / Off-road Setups
XTZ 125 — Part 1: Quando uma manutenção simples se transformou em um projeto →
The series documents the evolution of the project, technical decisions, experiments and results.
Backend Engineering · Applied AI · Cloud & Serverless · Product Engineering