"Finally, after having known some interesting languages... I can say that Rust is the only language I want to program with."
use std::Lang::Rust;
let software_developer = skill.map(|π¦| Rust(π¦)+π).collect();π Production & AI Systems β Live at @leonobitech
π Shipping in real-time βΒ
β
β
β
β
Everything below runs live in production on my own VPS under @leonobitech β my startup, where I ship production AI automation for SMB clients in LATAM.
The flagship is a complete agent stack for sales and customer service:
- π¬ WhatsApp sales agent that qualifies leads
- π Booking agent that schedules meetings and generates payment links
- π§© All orchestrated through custom MCP servers I built to connect Claude with client infrastructure: Odoo CRM Β· Qdrant (RAG) Β· n8n Β· Docker
| Repo | Stack | Role |
|---|---|---|
| π¨ frontend | Next.js Β· TypeScript | Customer-facing UI |
| βοΈ backend | Node Β· TypeScript Β· Hexagonal | Core API |
| π hardware | Rust Β· ESP32-C3 Β· ESP-IDF | IoT firmware with Secure Boot |
| ποΈ agents | Python Β· Realtime | Voice AI agents framework |
π Inside backend/repositories/ β The real microservices stack
The backend repo is a mono-repo with 15 production microservices. This is where the real engineering lives:
ποΈ Voice AI stack
| Service | Stack | Role |
|---|---|---|
| ποΈ voice-agent | Python Β· LiveKit | Realtime voice AI agent (LiveKit agents fork) |
| π voice-agent-3D | Python Β· LemonSlice | Voice agent with 3D avatar support |
| π¬ voice-spike | Python Β· Piper TTS | WebRTC voice spike (migrated from XTTS v2 β Piper TTS) |
| πΉ livekit | Docker | Production-hardened LiveKit setup |
| π£οΈ piper-tts | Python | Text-to-speech microservice |
π¦ Core services
| Service | Stack | Role |
|---|---|---|
| π¦ core-v2 | Rust | Complete Rust microservice foundation (+ blog series) |
| βοΈ core | Docker Β· Traefik | Infrastructure & auth core |
| π§ Qdrant | Docker | Vector database for RAG pipelines |
π€ AI & business integrations
| Service | Stack | Role |
|---|---|---|
| π 1-appointment-agent | Python | Appointment booking AI agent |
| π§© odoo-mcp | TypeScript | Custom MCP server for Odoo (appointments + billing) |
| π odoo | Python | Odoo 19 ERP configuration |
π οΈ Supporting infra
| Service | Stack | Role |
|---|---|---|
| π n8n | Docker | Workflow orchestration (1,084 nodes) |
| ποΈ baserow | Docker | No-code database backend (8 tables) |
| πΎ redis | Docker | Cache & queue backend |
| π² wa-signature-proxy | Node | WhatsApp webhook signature verification proxy |
| Repo | Description |
|---|---|
| π§© mcp-infrastructure | Custom MCP servers connecting Claude Code β VPS (SSH, n8n, Baserow, Notion) β operating 30 containers via natural language |
| π₯ fullstack-mcp-playground β | Fullstack AI agent template β Claude + microservices + modular tools from UI gallery |
| π°οΈ remote-connector-claude | Remote MCP Connector for Claude Desktop |
| Repo | Stack |
|---|---|
| π³ fullstack-infrastructure-blueprint | Docker Β· Traefik Β· HTTPS grid |
| π¨ fullstack-frontend-core | Next.js 15 Β· TS (ESM) Β· Docker |
| βοΈ fullstack-backend-core | Hexagonal API Β· Node Β· TS Β· ESM |
| Repo | Description |
|---|---|
| π¦π· l10n_ar_arca_edi | Odoo 19 Β· ARCA Electronic Invoicing (WSAA/WSFE) |
| π¦π· l10n-argentina | Odoo modules for Argentina |
ββββ VPS Β· 30 containers Β· Docker Compose Β· Traefik + Core (auth) βββββ
β β
β π€ AI / LLM Layer β
β β’ Custom MCP servers (Odoo Β· n8n Β· Baserow Β· SSH Β· Notion) β
β β’ Qdrant (vector DB for RAG pipelines) β
β β’ Voice AI agents (realtime conversational) β
β β’ n8n (1,084 nodes Β· workflow orchestration) β
β β
β πΌ Business Systems β
β β’ Odoo 19 (ERP + Argentina ARCA e-invoicing) β
β β’ Baserow (no-code DB Β· 8 tables) β
β β’ Chatwoot (customer messaging) β
β β
β π Product Stack β
β β’ Frontend (Next.js 15 Β· TS Β· Docker) β
β β’ Backend (Node Β· TS Β· Hexagonal Β· ESM) β
β β’ Hardware (ESP32-C3 Β· Rust Β· Secure Boot) β
β β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
- π Agent orchestration & pipelines β multi-step AI workflows where Claude routes decisions, tools, and data across the client's stack
- π§ͺ API systems β hexagonal architectures, gRPC services, async Rust backends powering the platform
- π¬ WhatsApp sales agent β qualifies leads automatically, live with SMB clients in LATAM
- π Booking agent β schedules meetings and generates payment links end-to-end
- π§© Custom MCP servers connecting Claude with Odoo CRM, Qdrant, n8n, and the VPS (30 containers operated via natural language)
- π§ RAG pipeline in production with Qdrant for context-aware AI responses
- π Odoo 19 with AFIP/ARCA e-invoicing β full fiscal compliance for Argentina
- ποΈ Realtime voice AI agents β conversational interfaces for customers
- π³ 30-container VPS β Docker Compose + Traefik + automatic HTTPS
- π IoT firmware (ESP32-C3, Rust) with Secure Boot for edge deployments
- π¦ Reusable blueprints published for fullstack + infra + MCP playgrounds
Electronic Engineer with 6+ years in software engineering (web, backend, API systems) and 3+ years deep in AI and LLM-based systems β currently orchestrating agents and pipelines, shipping real workflow automation powered by AI.
I work across the stack β Rust for embedded + backend (ESP32-C3, Axum, Tokio), TypeScript / Next.js for product, and custom MCP servers for the AI layer β bridging hardware, software, and AI end-to-end.
My biggest recent bet: launching Leonobitech, shipping production AI automation for SMB clients in LATAM. Details in the production section above β
AI / LLM / Automation
Languages & Core
Embedded & IoT
Backend & Infrastructure
Frontend
| Project | Description |
|---|---|
| paso-01-scaffold | Hands-on Rust embedded course with ESP32-C3 β from LED blink to concurrency and watchdog management (9 steps) |
| embedded-learning | ESP32 embedded systems learning repository β Rust & C |
| esp32-essential-training | ESP32 with ESP-IDF in C β foundation for embedded systems |
| Project | Description |
|---|---|
| actor-model-system-in-tokio | Actor model for order tracking with budget thresholds β concurrency with Tokio |
| asset-upload-in-Cloudinary | Asynchronous asset upload system to Cloudinary, built in Rust |
| axum-cookbook | Recipes with Tokio, Hyper, Serde, and Tower |
| grpc-tonic | gRPC service with Tonic and instrumentation via Autometrics |
| Project | Description |
|---|---|
| functional-programming-in-rust | Tutorials on fundamentals and advanced applications of functional programming in Rust |
| sha256-cracker | CLI sha256 cracker β Rust systems programming exercise |
| dining-philosophers | Classic concurrency problem implemented in Rust |
| Project | Description |
|---|---|
| blog_post_rust | Full-Stack web application with Rust and Leptos |
| superheroes | Authentication and access-control system with Redux β Next.js |


