AI Engineer · AI Consultant

I take AI to production,
from the LLM to the firmware.

I design and ship end-to-end AI solutions: from the language model and agents, to the microservices that serve them and the hardware they run on.

−94% event latency in a production system
9/9 modules migrated and verified on real hardware
3 generations of self-built assistants since 2023
20+ MCP servers built and operating
01about me

A computer engineer who doesn't stop at the prototype.

I've worked with AI since 2020 — from academic foundations (genetic algorithms, MCTS, decision trees) to building real agentic platforms: voice and generative-AI assistants on microservice architectures, STT/TTS/NLU pipelines, custom MCP servers, and LLM observability with Langfuse, Grafana and Prometheus.

I don't stop at the demo: I orchestrate, monitor and operate. Today I apply AI to engineering and industrial workflows, including environments where AI lives alongside hardware.

Track record
  1. 2020 AI foundations at university: genetic algorithms, MCTS, decision trees.
  2. 2023 nuka — 1st self-built assistant generation: GPT, Whisper, TTS and RAG, in the first wave of generative AI.
  3. 2025 puertocho-assistant — 2nd generation: microservices and an E2E voice pipeline with model voting (MoE).
  4. now tony — 3rd generation: an agentic platform with 20+ MCPs, continuous evaluation and its own hardware. Plus applied-AI consulting in industrial environments.
02how i help companies

From AI pilots that dazzle in a demo, to systems that actually work in production.

/agents

Agent deployment & governance

Your company wants to use AI agents, but the jump from notebook to production is where almost everyone gets stuck. I close that gap.

  • Agent and MCP integration architecture
  • LLM observability and tracing (Langfuse)
  • Cost, latency and reliability under control
/hardware

AI in hardware environments

I integrate AI into industrial and device workflows: from backend to firmware. If your product has electronics, you don't need two vendors.

  • AI applied to engineering and industrial processes
  • Software ↔ device bridge (CAN, industrial protocols)
  • Microservices and distributed architectures
/content

Content & generative AI

Systems to generate and manage content with AI — from scripts to publishing — built with engineering rigor, not just loose prompts.

  • Custom generative-AI pipelines
  • Content agents and automation
  • Generative-AI adoption advisory
03selected projects

Personal AI R&D that underpins my professional practice. The code is open.

04contact

Got an AI pilot you want to take to production?

Let's talk. Tell me the problem and I'll be straight about whether I can help and how.

info@antoniopuerto.com