Computer Vision & AI Engineer · Madrid, Spain
Real-time vision systems for industrial robotics · LLM agents and MCP
AI and software engineer building real-time vision systems for industrial robotics, from model training to GPU inference and robot integration. Previously: MCP servers and LLM agents for enterprise automation, and a game shipped on Steam. Computer Science graduate from the University of Havana.
| Project | Description | Links |
|---|---|---|
| Industrial Vision System | Current work: real-time detection, pose estimation and OCR on TensorRT | Private |
| simple-mcp-server | MCP server in Python over Streamable HTTP | Repo |
| fastchat-mcp | Python client to chat with MCP servers in natural language | Repo · PyPI |
| MCP Utils | Index of my MCP clients, servers, OAuth and templates | Repo |
| Arcane Ascent | Deck-building roguelike in Unity/C#, released on Steam | Steam · Repo |
Experience
| Period | Role | Company |
|---|---|---|
| 2026 – Present | Computer Vision & AI Engineer | Promart Automatizaciones · Madrid, Spain |
| 2025 – 2026 | AI Engineer (LLM Agents & MCP) | Cogitia · Remote |
| 2024 – 2025 | Game Developer | Word Games · Remote |
| 2024 | NLP & Backend Developer | Devman · Remote |
Tech Stack
| Area | Technologies |
|---|---|
| Computer Vision | Object detection, pose estimation, OCR, PyTorch, TensorFlow, ONNX Runtime, TensorRT, Optuna |
| LLM Systems | MCP, LangGraph, LangChain, RAG, NLP, OAuth 2.1 |
| Backend | Python, FastAPI, .NET, Docker, Git |
| Languages | Python, C#, C/C++, JavaScript/TypeScript |
| Other | React, PyQt6, Unity |
Research
| Paper | Summary | Links |
|---|---|---|
| Image Retrieval with SAM and CLIP (B.Sc. thesis) | Text-to-image retrieval combining SAM segmentation with CLIP embeddings | Paper · Code |
| Music Genre Classification | CNN and RNN models on GTZAN, 75–80% accuracy | Paper · Code |
