Enmanuel
Molina.

I architect high-performance systems across the full stack — Rust/Go/gRPC backends, Node.js/TS microservices, Flutter mobile apps — with a specialized focus on Local-First AI.

Contact void-stack.dev
  • 11+ years experience
  • 42 MCP tools
  • ~70% token reduction
  • 3× faster delivery

01. about

I build complete systems
and make them intelligent.

Beyond API wrappers — actual inference pipelines running locally with GPU optimization. From satellite imagery analysis to document extraction, I integrate AI at the infrastructure level.

Creator of void-stack , an open-source MCP server in Rust used daily in production development workflows.

technical stack

Languages
Rust · Go · TypeScript · Python · Dart
Backend
Node.js · gRPC · NestJS · Express
Mobile
Flutter (iOS · Android · Desktop)
Frontend
Next.js · React · Astro
AI/Local
Ollama · LLaVA · SAM · ESRGAN · MCP
Cloud
AWS · DigitalOcean · Cloudflare · Terraform
Databases
PostgreSQL · Redis · Supabase · Meilisearch

02. projects

Selected work.

  1. 01

    void-stack

    OPEN SOURCE

    MCP server in Rust — 42 tools for AI-assisted development

    Open-source dev tool that connects Claude Code and Cursor to your full stack via MCP protocol. Hybrid semantic + structural analysis, service management, security audit, live call graphs.

    ~70% token reduction in real-world use

    • Rust
    • Tauri v2
    • gRPC
    • tree-sitter
    • MCP Protocol
    void-stack.dev
  2. 02

    iunci

    PRODUCTION

    POS & business management — Flutter offline-first, real clients

    Business management platform with serialized inventory, offline-first sync, Google Drive backup and financing integration. Multi-tenant marketplace with instant search.

    Go/gRPC backend · 77.8% test coverage · production

    • Flutter
    • Go
    • gRPC
    • Node.js
    • TypeScript
    • PostgreSQL
    iunci.app
  3. 03

    Local AI Pipelines

    PERSONAL

    Satellite & document analysis with local models

    Production pipelines running LLaVA, SAM, ESRGAN and ControlNet locally via Ollama with full GPU optimization — Flash Attention, quantized KV Cache, VRAM management between model phases.

    GPU-optimized local inference — Flash Attention, quantized KV Cache

    • Python
    • CUDA/CuPy
    • Ollama
    • FastAPI
    • Sentinel-2
    • SRTM

03. experience

A decade-plus building software.

  1. Jun 2025 → Present ● now

    Backend Engineer SquadMakers (Remote)

    Developing microservices architecture for Spain-based client.

    • Node.js
    • TypeScript
    • PostgreSQL
    • Docker

  2. Jul 2024 → May 2025

    Back End Developer Lodgerin (Remote · Spain)

    Stripe payments, Signaturit digital contracts, Node.js APIs, Ruby on Rails legacy migration.

  3. Apr 2022 → May 2024

    Software Engineer Acid Labs / Cencosud (Remote)

    AWS infrastructure with Terraform (Lambda, Fargate, S3, SQS, DynamoDB), NestJS + TypeScript, CI/CD GitLab + GitHub Actions.

  4. 2014 → 2021

    Software Engineer Freelance & own projects (Remote · Panama · Spain)

    Go, Node.js, Flutter, PHP, Java. Clients in Panama and Spain.

04. contact

Let's build
something.

Available for remote work worldwide. UTC-4 — compatible with Spain, Europe, and Latin America.