I'm Michel Freitas, an 18-year-old Software Engineer at R2 Ventures, based in São Paulo, Brazil.
I started programming when I was 12 and have been building real-world products ever since — from full-stack platforms and internal enterprise systems to AI-powered applications, embedded software, cloud tools and desktop apps.
Currently, I work mainly with Go, Node.js, NestJS, AWS and Terraform, with a strong focus on backend engineering, cloud infrastructure, data/AI systems and scalable software architecture.
const michel = {
age: 18,
role: "Software Engineer",
company: "R2 Ventures",
location: "São Paulo, Brazil",
mainStack: ["Go", "Node.js", "NestJS", "AWS", "Terraform"],
interests: [
"Backend Engineering",
"Cloud Infrastructure",
"Artificial Intelligence",
"Distributed Systems",
"Developer Tools",
"Desktop Apps",
],
currentlyBuilding: ["MildStack", "Flyra", "Cloud-native tools"],
};Flyra is a complete system using drones equipped with artificial intelligence to detect and help prevent wildfires.
I worked across the entire product: AI training, embedded systems, backend, desktop/mobile applications, real-time drone integration and team management during the final technical school project.
Website: flyra.michelfreitas.com
- AI model for fire detection
- Local embedded server running on a Raspberry Pi
- Central backend for teams, users, drones and organizations
- Real-time drone camera and sensor data transmission
- Desktop/mobile app for monitoring, reports and alerts
- Project architecture, development coordination and team management
| Repository | Description |
|---|---|
| flyra-local-server | Local server running on a Raspberry Pi. Executes the AI model, collects drone sensor data, streams video and sends information to the central backend. |
| flyra-central-server | Central backend responsible for teams, users, drones, organizations and more. Built with DDD, NestJS, TypeScript and Prisma. |
| flyra-desktop-tauri | Multiplatform app built with Tauri + Rust, with support for desktop, mobile and web. Used for real-time camera visualization, drone communication, reports and alerts. |
MildStack is a free, lightweight and open-source alternative to LocalStack.
The idea is simple: run your own AWS-like environment locally with a fast, native Go runtime — without Docker, without cloud accounts and without heavy setup.
Website: mildstack.dev Repository: github.com/michasdev/mildstack
- 100% Go-based core
- No Docker required
- CLI for running local cloud services
- Desktop app for browsing and managing resources
- Resource Browser for local AWS-like services
- Built to be lightweight, fast and developer-friendly
- S3
- DynamoDB
- SQS
- SNS
- Lambda
- EventBridge
- CloudWatch Logs
- And more
- eFocus — robust e-commerce platform for textile sales, built with Next.js and TypeScript.
- RAG System with AI — semantic search and AI-powered fabric recommendation system using NestJS, TypeScript, LangChain, Ollama and MongoDB Vector Search.
- Internal enterprise systems — development and maintenance of full-stack systems, APIs, CI/CD pipelines and integrations.
- Freitas UI — experimental React component library focused on modern, animated and editable components.
mindmap
root((Michel Freitas))
Backend
Go
Node.js
NestJS
DDD
APIs
Cloud
AWS
Terraform
Infrastructure
Local cloud tools
AI
RAG Systems
Computer Vision
TensorFlow
Embeddings
Apps
Tauri
Desktop
Mobile
Web
Systems
Real-time communication
Embedded software
Developer tools
- Building cloud and infrastructure tools
- Improving my Go and AWS ecosystem knowledge
- Creating high-performance backend systems
- Exploring AI applied to real-world products
- Developing tools that make developers faster
