Senior Software Engineer with 10+ years building backend services and full-stack products for global fintech and enterprise clients. These days I build agentic AI systems: LLM orchestration with guardrails, observability and output validation. On the side, I tinker with local-first AI on consumer GPUs and Apple Silicon.
- 🤖 Currently: agentic delivery workflows (research → plan → implement → review) on Claude and OpenAI
- 🖥️ Side quests: on-device inference with Ollama, MLX and ComfyUI; macOS tooling in Swift
- ☕ Background: Java/Spring microservices, API performance and legacy modernization
| Project | What it is | Stack |
|---|---|---|
| riftway | Run Windows games and installers on macOS with a double click: its own Wine runtime, library, URL bridges and compatibility profiles | Swift |
| fmodel-mac | Unofficial macOS (Apple Silicon) port of FModel, built on CUE4Parse | C#, Avalonia |
| ComfyUI-Model-Installer | Installs the models a ComfyUI workflow needs | Python |
Other things I've built (private repos)
- Monarca Studio: local-first desktop app for on-device image and video generation (Electron + FastAPI)
- LAICA: AI-assisted IDE with built-in Git Flow guidance (Electron, React, TypeScript)
- ComfyUI API Manager: exposes ComfyUI workflows as a REST API (FastAPI, React, SQLite)
- VideoForge: AI video generation backends for CUDA and MLX
- CanYouRunAI: checks whether a machine can run a given AI model locally
- AI / LLM: Claude & OpenAI APIs · agentic workflows · prompt & context engineering · Ollama · MLX · ComfyUI
- Languages: Java · Python · Go · TypeScript · Swift
- Backend: Spring Boot · FastAPI · Node.js · Kafka · RabbitMQ · PostgreSQL · Redis
- Frontend & desktop: React · Vite · Tailwind · Electron · SwiftUI
- Cloud & DevOps: AWS · Azure · Docker · Terraform
📍 Mexico · 🗣️ Spanish / English